[transformers] Token indices sequence length is longer than the specified maximum sequence length for this model (317092 > 262144). Running this sequence through the model will result in indexing errors
[i8] basket: 20 prompts x 16384 tok, domains ['code', 'longctx', 'math', 'sci', 'wiki']
compile-cache MISS: xgrammar — xgrammar compiled grammars (re)compiles this boot [/cache/xgrammar]
default-pool residency snapshot returned no live ranges — the unpooled-weight fold cannot see what it would move
Boot: building KV-cache profile — still running (30s). This phase reports nothing until it completes; it is progressing while this line advances. If it stops advancing, the phase is stuck.
[TKV autotune-provenance] fp=57911e26f4e47f74 regime=cache-hit cells=8 sweep_src=2772549a0ef713db tile_tokens=4 num_splits=32,64,128,256 min_blocks_per_sm=0 resolved=0 held=0 inherited=0
driver.modules_loaded baseline: SEEDED at 153092096 B (0.143 GiB) → /root/.cache/arbi-serve/budget-cache/84169d322e956268.modules.json. This boot ran UNGATED — it booked its own bracketed growth, so an unregistered pool or a raw cudaMalloc would be inside that number rather than on driver.residual. Every later boot at this configuration is held to it. Expected ONCE per configuration; if it repeats, the budget cache is not persisting (point ARBI_SERVE_BUDGET_CACHE_DIR at a durable volume) and the guard is inert.
/opt/venv/lib/python3.12/site-packages/torch/__init__.py:1172: FutureWarning: `torch.distributed.reduce_op` is deprecated, please use `torch.distributed.ReduceOp` instead
  return isinstance(obj, torch.Tensor)
[i8] trellis->path map: 401 linears
[i8] layer trace: 64 decoder layers hooked
JIT compile AFTER serving-ready [cpp_ext]: exl3_i8_gemm fresh ninja/nvcc build — a live request paid this compile's latency. This is a boot-warmup coverage gap: extend warmup to pre-compile this kernel/specialization. Counter jit_compile_serving (must-not-fire) at GET /v1/admin/flag_truth.
[i8] prefix cache: server default resolved=True, per-request cache_enabled=False; any cache hit below is refused
driver serving-growth OVER RESERVE: the driver holds 2 MiB more than it did when the boot brackets closed, against the 0 MiB the serving floor's transient.serving_step.driver_growth term held for it. Those bytes are in no pool and no allocator counter -- the live card reaches them as driver.modules_serving and the boot ledger only as driver.residual -- and the KV layout is frozen, so the overage comes out of the free VRAM the floor is holding for one in-flight step (the verify tail, the DFlash context-assemble) and those reserves are the ones that fail first. The reading is on record for this configuration, so the next boot holds it; this process serves the rest of its life short by the overage.
activation admission RE-ARMED (driver residency over reserve): the step budget FELL to 5741 MiB, -642 MiB on the 6383 MiB it was enforcing and 6383 MiB the boot reading armed. The widest slate admission will build is narrower from here — narrower prefill chunks and deferred rows instead of a forward that OOMs after the layout freeze.
[i8] warmup done
  [0] code           REF      16384tok  wall=  5.01s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [0] code           REF2     16384tok  wall=  5.00s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [0] code           STD      16384tok  wall=  5.26s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [0] code           SIMROW   16384tok  wall=  7.51s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [0] code           INT8     16384tok  wall=  4.83s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [1] code           REF      16384tok  wall=  5.03s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [1] code           REF2     16384tok  wall=  5.04s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [1] code           STD      16384tok  wall=  5.30s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [1] code           SIMROW   16384tok  wall=  7.54s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [1] code           INT8     16384tok  wall=  4.84s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [2] code           REF      16384tok  wall=  5.05s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [2] code           REF2     16384tok  wall=  5.07s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [2] code           STD      16384tok  wall=  5.31s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [2] code           SIMROW   16384tok  wall=  7.54s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [2] code           INT8     16384tok  wall=  4.84s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [3] code           REF      16384tok  wall=  5.06s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [3] code           REF2     16384tok  wall=  5.06s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [3] code           STD      16384tok  wall=  5.33s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [3] code           SIMROW   16384tok  wall=  7.55s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [3] code           INT8     16384tok  wall=  4.84s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [4] wiki           REF      16384tok  wall=  5.07s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [4] wiki           REF2     16384tok  wall=  5.07s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [4] wiki           STD      16384tok  wall=  5.35s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [4] wiki           SIMROW   16384tok  wall=  7.56s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [4] wiki           INT8     16384tok  wall=  4.84s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [5] wiki           REF      16384tok  wall=  5.09s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [5] wiki           REF2     16384tok  wall=  5.08s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [5] wiki           STD      16384tok  wall=  5.33s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [5] wiki           SIMROW   16384tok  wall=  7.56s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [5] wiki           INT8     16384tok  wall=  4.84s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [6] wiki           REF      16384tok  wall=  5.08s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [6] wiki           REF2     16384tok  wall=  5.09s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [6] wiki           STD      16384tok  wall=  5.34s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [6] wiki           SIMROW   16384tok  wall=  7.63s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [6] wiki           INT8     16384tok  wall=  4.85s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [7] wiki           REF      16384tok  wall=  5.09s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [7] wiki           REF2     16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [7] wiki           STD      16384tok  wall=  5.33s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [7] wiki           SIMROW   16384tok  wall=  7.57s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [7] wiki           INT8     16384tok  wall=  4.90s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [8] math           REF      16384tok  wall=  5.09s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [8] math           REF2     16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [8] math           STD      16384tok  wall=  5.33s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [8] math           SIMROW   16384tok  wall=  7.57s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [8] math           INT8     16384tok  wall=  4.85s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [9] math           REF      16384tok  wall=  5.09s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [9] math           REF2     16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [9] math           STD      16384tok  wall=  5.35s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [9] math           SIMROW   16384tok  wall=  7.57s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [9] math           INT8     16384tok  wall=  4.86s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [10] math           REF      16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [10] math           REF2     16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [10] math           STD      16384tok  wall=  5.34s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [10] math           SIMROW   16384tok  wall=  7.58s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [10] math           INT8     16384tok  wall=  4.86s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [11] math           REF      16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [11] math           REF2     16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [11] math           STD      16384tok  wall=  5.34s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [11] math           SIMROW   16384tok  wall=  7.57s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [11] math           INT8     16384tok  wall=  4.86s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [12] longctx        REF      16384tok  wall=  5.09s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [12] longctx        REF2     16384tok  wall=  5.09s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [12] longctx        STD      16384tok  wall=  5.34s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [12] longctx        SIMROW   16384tok  wall=  7.57s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [12] longctx        INT8     16384tok  wall=  4.86s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [13] longctx        REF      16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [13] longctx        REF2     16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [13] longctx        STD      16384tok  wall=  5.34s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [13] longctx        SIMROW   16384tok  wall=  7.57s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [13] longctx        INT8     16384tok  wall=  4.86s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [14] longctx        REF      16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [14] longctx        REF2     16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [14] longctx        STD      16384tok  wall=  5.35s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [14] longctx        SIMROW   16384tok  wall=  7.58s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [14] longctx        INT8     16384tok  wall=  4.86s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [15] longctx        REF      16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [15] longctx        REF2     16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [15] longctx        STD      16384tok  wall=  5.34s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [15] longctx        SIMROW   16384tok  wall=  7.58s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [15] longctx        INT8     16384tok  wall=  4.86s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [16] sci            REF      16384tok  wall=  5.25s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [16] sci            REF2     16384tok  wall=  5.22s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [16] sci            STD      16384tok  wall=  5.50s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [16] sci            SIMROW   16384tok  wall=  7.73s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [16] sci            INT8     16384tok  wall=  5.01s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [17] sci            REF      16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [17] sci            REF2     16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [17] sci            STD      16384tok  wall=  5.35s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [17] sci            SIMROW   16384tok  wall=  7.57s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [17] sci            INT8     16384tok  wall=  4.85s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [18] sci            REF      16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [18] sci            REF2     16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [18] sci            STD      16384tok  wall=  5.35s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [18] sci            SIMROW   16384tok  wall=  7.58s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [18] sci            INT8     16384tok  wall=  4.86s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [19] sci            REF      16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [19] sci            REF2     16384tok  wall=  5.10s  finish=length  leg=3200F/0S acc=3200x16/0x32
  [19] sci            STD      16384tok  wall=  5.35s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [19] sci            SIMROW   16384tok  wall=  7.58s  finish=length  leg=0F/3200S acc=3200x16/0x32
  [19] sci            INT8     16384tok  wall=  4.85s  finish=length  leg=0F/3200S acc=3200x16/0x32

[i8] kernel census: int8 calls=64000 rows=131072000 reconstructs SKIPPED=64000 fallback (non-4bpw) calls=0
[i8] shapes served by the kernel (K,N)->calls: {(5120, 1024): 5120, (5120, 6144): 7680, (5120, 10240): 7680, (5120, 12288): 2560, (5120, 17408): 20480, (6144, 5120): 10240, (17408, 5120): 10240}
[i8] excluded projections: none  excluded calls=0

  linear class                                           served   excl  int8 vs fp32 legB vs fp32 int8 vs legB
  model.layers.0.linear_attn.in_proj_qkv                    161      0     4.827e-03    1.713e-03    5.132e-03
  model.layers.0.linear_attn.in_proj_z                      161      0     6.316e-03    1.129e-03    6.418e-03
  model.layers.0.linear_attn.out_proj                       161      0     1.249e-02    2.189e-03    1.268e-02
  model.layers.0.mlp.down_proj                              161      0     2.122e-02    2.609e-03    2.137e-02
  model.layers.0.mlp.gate_proj                              161      0     9.121e-03    3.413e-03    9.739e-03
  model.layers.0.mlp.up_proj                                161      0     9.082e-03    3.418e-03    9.702e-03
  model.layers.1.linear_attn.in_proj_qkv                    161      0     1.186e-02    2.166e-03    1.206e-02
  model.layers.1.linear_attn.in_proj_z                      161      0     1.361e-02    1.840e-03    1.373e-02
  model.layers.1.linear_attn.out_proj                       161      0     1.604e-02    2.253e-03    1.619e-02
  model.layers.1.mlp.down_proj                              161      0     1.345e-02    2.322e-03    1.364e-02
  model.layers.1.mlp.gate_proj                              161      0     9.700e-03    3.459e-03    1.030e-02
  model.layers.1.mlp.up_proj                                161      0     1.015e-02    3.488e-03    1.073e-02
  model.layers.10.linear_attn.in_proj_qkv                   161      0     1.304e-02    2.256e-03    1.323e-02
  model.layers.10.linear_attn.in_proj_z                     161      0     1.298e-02    1.838e-03    1.310e-02
  model.layers.10.linear_attn.out_proj                      161      0     1.498e-02    2.124e-03    1.513e-02
  model.layers.10.mlp.down_proj                             161      0     1.268e-02    2.347e-03    1.290e-02
  model.layers.10.mlp.gate_proj                             161      0     1.035e-02    3.578e-03    1.095e-02
  model.layers.10.mlp.up_proj                               161      0     1.075e-02    3.590e-03    1.134e-02
  model.layers.11.mlp.down_proj                             161      0     1.337e-02    2.508e-03    1.360e-02
  model.layers.11.mlp.gate_proj                             161      0     1.027e-02    3.508e-03    1.085e-02
  model.layers.11.mlp.up_proj                               161      0     1.087e-02    3.574e-03    1.144e-02
  model.layers.11.self_attn.k_proj                          161      0     1.329e-02    2.697e-03    1.356e-02
  model.layers.11.self_attn.o_proj                          161      0     1.303e-02    2.370e-03    1.324e-02
  model.layers.11.self_attn.q_proj                          161      0     1.173e-02    2.022e-03    1.190e-02
  model.layers.11.self_attn.v_proj                          161      0     1.036e-02    2.280e-03    1.061e-02
  model.layers.12.linear_attn.in_proj_qkv                   161      0     1.292e-02    2.245e-03    1.311e-02
  model.layers.12.linear_attn.in_proj_z                     161      0     1.366e-02    1.891e-03    1.379e-02
  model.layers.12.linear_attn.out_proj                      161      0     1.586e-02    2.138e-03    1.600e-02
  model.layers.12.mlp.down_proj                             161      0     1.316e-02    2.553e-03    1.341e-02
  model.layers.12.mlp.gate_proj                             161      0     1.003e-02    3.475e-03    1.061e-02
  model.layers.12.mlp.up_proj                               161      0     1.073e-02    3.555e-03    1.130e-02
  model.layers.13.linear_attn.in_proj_qkv                   161      0     1.292e-02    2.241e-03    1.311e-02
  model.layers.13.linear_attn.in_proj_z                     161      0     1.418e-02    1.971e-03    1.432e-02
  model.layers.13.linear_attn.out_proj                      161      0     1.683e-02    2.176e-03    1.697e-02
  model.layers.13.mlp.down_proj                             161      0     1.372e-02    2.566e-03    1.396e-02
  model.layers.13.mlp.gate_proj                             161      0     1.006e-02    3.483e-03    1.065e-02
  model.layers.13.mlp.up_proj                               161      0     1.071e-02    3.558e-03    1.128e-02
  model.layers.14.linear_attn.in_proj_qkv                   161      0     1.271e-02    2.230e-03    1.290e-02
  model.layers.14.linear_attn.in_proj_z                     161      0     1.336e-02    1.881e-03    1.349e-02
  model.layers.14.linear_attn.out_proj                      161      0     1.586e-02    2.216e-03    1.601e-02
  model.layers.14.mlp.down_proj                             161      0     1.299e-02    2.463e-03    1.322e-02
  model.layers.14.mlp.gate_proj                             161      0     9.956e-03    3.503e-03    1.056e-02
  model.layers.14.mlp.up_proj                               161      0     1.064e-02    3.566e-03    1.122e-02
  model.layers.15.mlp.down_proj                             161      0     1.350e-02    2.531e-03    1.374e-02
  model.layers.15.mlp.gate_proj                             161      0     1.010e-02    3.487e-03    1.069e-02
  model.layers.15.mlp.up_proj                               161      0     1.068e-02    3.563e-03    1.126e-02
  model.layers.15.self_attn.k_proj                          161      0     1.453e-02    2.890e-03    1.481e-02
  model.layers.15.self_attn.o_proj                          161      0     1.398e-02    2.119e-03    1.414e-02
  model.layers.15.self_attn.q_proj                          161      0     1.300e-02    2.088e-03    1.317e-02
  model.layers.15.self_attn.v_proj                          161      0     9.817e-03    2.224e-03    1.007e-02
  model.layers.16.linear_attn.in_proj_qkv                   161      0     1.290e-02    2.245e-03    1.309e-02
  model.layers.16.linear_attn.in_proj_z                     161      0     1.355e-02    1.897e-03    1.368e-02
  model.layers.16.linear_attn.out_proj                      161      0     1.636e-02    2.282e-03    1.652e-02
  model.layers.16.mlp.down_proj                             161      0     1.300e-02    2.516e-03    1.324e-02
  model.layers.16.mlp.gate_proj                             161      0     1.005e-02    3.471e-03    1.063e-02
  model.layers.16.mlp.up_proj                               161      0     1.070e-02    3.563e-03    1.128e-02
  model.layers.17.linear_attn.in_proj_qkv                   161      0     1.288e-02    2.224e-03    1.307e-02
  model.layers.17.linear_attn.in_proj_z                     161      0     1.351e-02    1.896e-03    1.364e-02
  model.layers.17.linear_attn.out_proj                      161      0     1.776e-02    2.244e-03    1.790e-02
  model.layers.17.mlp.down_proj                             161      0     1.380e-02    2.453e-03    1.401e-02
  model.layers.17.mlp.gate_proj                             161      0     1.009e-02    3.470e-03    1.067e-02
  model.layers.17.mlp.up_proj                               161      0     1.072e-02    3.567e-03    1.129e-02
  model.layers.18.linear_attn.in_proj_qkv                   161      0     1.206e-02    2.187e-03    1.226e-02
  model.layers.18.linear_attn.in_proj_z                     161      0     1.264e-02    1.787e-03    1.276e-02
  model.layers.18.linear_attn.out_proj                      161      0     1.526e-02    2.229e-03    1.542e-02
  model.layers.18.mlp.down_proj                             161      0     1.458e-02    2.426e-03    1.478e-02
  model.layers.18.mlp.gate_proj                             161      0     1.023e-02    3.425e-03    1.079e-02
  model.layers.18.mlp.up_proj                               161      0     1.059e-02    3.530e-03    1.116e-02
  model.layers.19.mlp.down_proj                             161      0     1.587e-02    2.542e-03    1.608e-02
  model.layers.19.mlp.gate_proj                             161      0     1.007e-02    3.395e-03    1.063e-02
  model.layers.19.mlp.up_proj                               161      0     1.096e-02    3.578e-03    1.153e-02
  model.layers.19.self_attn.k_proj                          161      0     1.411e-02    2.794e-03    1.438e-02
  model.layers.19.self_attn.o_proj                          161      0     1.813e-02    2.164e-03    1.826e-02
  model.layers.19.self_attn.q_proj                          161      0     9.504e-03    1.929e-03    9.697e-03
  model.layers.19.self_attn.v_proj                          161      0     1.045e-02    2.357e-03    1.071e-02
  model.layers.2.linear_attn.in_proj_qkv                    161      0     1.248e-02    2.213e-03    1.268e-02
  model.layers.2.linear_attn.in_proj_z                      161      0     1.399e-02    1.926e-03    1.412e-02
  model.layers.2.linear_attn.out_proj                       161      0     1.795e-02    2.320e-03    1.810e-02
  model.layers.2.mlp.down_proj                              161      0     1.706e-02    2.375e-03    1.722e-02
  model.layers.2.mlp.gate_proj                              161      0     1.029e-02    3.535e-03    1.088e-02
  model.layers.2.mlp.up_proj                                161      0     1.062e-02    3.558e-03    1.120e-02
  model.layers.20.linear_attn.in_proj_qkv                   161      0     1.214e-02    2.136e-03    1.233e-02
  model.layers.20.linear_attn.in_proj_z                     161      0     1.271e-02    1.793e-03    1.284e-02
  model.layers.20.linear_attn.out_proj                      161      0     1.578e-02    2.224e-03    1.594e-02
  model.layers.20.mlp.down_proj                             161      0     1.593e-02    2.583e-03    1.614e-02
  model.layers.20.mlp.gate_proj                             161      0     1.008e-02    3.408e-03    1.064e-02
  model.layers.20.mlp.up_proj                               161      0     1.126e-02    3.603e-03    1.183e-02
  model.layers.21.linear_attn.in_proj_qkv                   161      0     1.199e-02    2.201e-03    1.219e-02
  model.layers.21.linear_attn.in_proj_z                     161      0     1.162e-02    1.672e-03    1.174e-02
  model.layers.21.linear_attn.out_proj                      161      0     1.702e-02    2.231e-03    1.716e-02
  model.layers.21.mlp.down_proj                             161      0     1.721e-02    2.588e-03    1.740e-02
  model.layers.21.mlp.gate_proj                             161      0     1.010e-02    3.421e-03    1.066e-02
  model.layers.21.mlp.up_proj                               161      0     1.122e-02    3.605e-03    1.178e-02
  model.layers.22.linear_attn.in_proj_qkv                   161      0     1.201e-02    2.241e-03    1.222e-02
  model.layers.22.linear_attn.in_proj_z                     161      0     1.204e-02    1.712e-03    1.216e-02
  model.layers.22.linear_attn.out_proj                      161      0     1.579e-02    2.305e-03    1.595e-02
  model.layers.22.mlp.down_proj                             161      0     1.513e-02    2.552e-03    1.533e-02
  model.layers.22.mlp.gate_proj                             161      0     1.052e-02    3.137e-03    1.098e-02
  model.layers.22.mlp.up_proj                               161      0     1.142e-02    3.622e-03    1.197e-02
  model.layers.23.mlp.down_proj                             161      0     1.569e-02    2.559e-03    1.589e-02
  model.layers.23.mlp.gate_proj                             161      0     1.009e-02    3.369e-03    1.064e-02
  model.layers.23.mlp.up_proj                               161      0     1.141e-02    3.621e-03    1.197e-02
  model.layers.23.self_attn.k_proj                          161      0     1.378e-02    2.671e-03    1.404e-02
  model.layers.23.self_attn.o_proj                          161      0     1.682e-02    2.421e-03    1.699e-02
  model.layers.23.self_attn.q_proj                          161      0     9.478e-03    1.963e-03    9.680e-03
  model.layers.23.self_attn.v_proj                          161      0     1.034e-02    2.327e-03    1.060e-02
  model.layers.24.linear_attn.in_proj_qkv                   161      0     1.188e-02    2.253e-03    1.209e-02
  model.layers.24.linear_attn.in_proj_z                     161      0     1.221e-02    1.730e-03    1.234e-02
  model.layers.24.linear_attn.out_proj                      161      0     1.603e-02    2.268e-03    1.619e-02
  model.layers.24.mlp.down_proj                             161      0     1.458e-02    2.617e-03    1.481e-02
  model.layers.24.mlp.gate_proj                             161      0     1.026e-02    3.317e-03    1.078e-02
  model.layers.24.mlp.up_proj                               161      0     1.144e-02    3.613e-03    1.199e-02
  model.layers.25.linear_attn.in_proj_qkv                   161      0     1.142e-02    2.212e-03    1.163e-02
  model.layers.25.linear_attn.in_proj_z                     161      0     1.279e-02    1.788e-03    1.291e-02
  model.layers.25.linear_attn.out_proj                      161      0     1.473e-02    2.175e-03    1.489e-02
  model.layers.25.mlp.down_proj                             161      0     1.393e-02    2.554e-03    1.416e-02
  model.layers.25.mlp.gate_proj                             161      0     1.054e-02    3.322e-03    1.105e-02
  model.layers.25.mlp.up_proj                               161      0     1.130e-02    3.600e-03    1.185e-02
  model.layers.26.linear_attn.in_proj_qkv                   161      0     1.149e-02    2.236e-03    1.171e-02
  model.layers.26.linear_attn.in_proj_z                     161      0     1.260e-02    1.796e-03    1.272e-02
  model.layers.26.linear_attn.out_proj                      161      0     1.492e-02    2.343e-03    1.510e-02
  model.layers.26.mlp.down_proj                             161      0     1.270e-02    2.468e-03    1.294e-02
  model.layers.26.mlp.gate_proj                             161      0     1.112e-02    3.297e-03    1.160e-02
  model.layers.26.mlp.up_proj                               161      0     1.086e-02    3.552e-03    1.143e-02
  model.layers.27.mlp.down_proj                             161      0     1.335e-02    2.575e-03    1.359e-02
  model.layers.27.mlp.gate_proj                             161      0     1.079e-02    3.204e-03    1.126e-02
  model.layers.27.mlp.up_proj                               161      0     1.090e-02    3.555e-03    1.146e-02
  model.layers.27.self_attn.k_proj                          161      0     1.215e-02    2.498e-03    1.240e-02
  model.layers.27.self_attn.o_proj                          161      0     1.707e-02    2.745e-03    1.729e-02
  model.layers.27.self_attn.q_proj                          161      0     1.020e-02    2.062e-03    1.041e-02
  model.layers.27.self_attn.v_proj                          161      0     9.850e-03    2.252e-03    1.011e-02
  model.layers.28.linear_attn.in_proj_qkv                   161      0     1.177e-02    2.217e-03    1.198e-02
  model.layers.28.linear_attn.in_proj_z                     161      0     1.287e-02    1.817e-03    1.299e-02
  model.layers.28.linear_attn.out_proj                      161      0     1.479e-02    2.297e-03    1.497e-02
  model.layers.28.mlp.down_proj                             161      0     1.301e-02    2.600e-03    1.326e-02
  model.layers.28.mlp.gate_proj                             161      0     1.104e-02    3.175e-03    1.149e-02
  model.layers.28.mlp.up_proj                               161      0     1.084e-02    3.537e-03    1.140e-02
  model.layers.29.linear_attn.in_proj_qkv                   161      0     1.204e-02    2.246e-03    1.225e-02
  model.layers.29.linear_attn.in_proj_z                     161      0     1.397e-02    1.967e-03    1.411e-02
  model.layers.29.linear_attn.out_proj                      161      0     1.655e-02    2.350e-03    1.671e-02
  model.layers.29.mlp.down_proj                             161      0     1.348e-02    2.623e-03    1.373e-02
  model.layers.29.mlp.gate_proj                             161      0     1.114e-02    3.188e-03    1.159e-02
  model.layers.29.mlp.up_proj                               161      0     1.081e-02    3.536e-03    1.137e-02
  model.layers.3.mlp.down_proj                              161      0     1.418e-02    2.577e-03    1.441e-02
  model.layers.3.mlp.gate_proj                              161      0     1.018e-02    3.528e-03    1.077e-02
  model.layers.3.mlp.up_proj                                161      0     1.101e-02    3.601e-03    1.158e-02
  model.layers.3.self_attn.k_proj                           161      0     1.249e-02    2.587e-03    1.275e-02
  model.layers.3.self_attn.o_proj                           161      0     1.494e-02    2.175e-03    1.510e-02
  model.layers.3.self_attn.q_proj                           161      0     1.169e-02    2.033e-03    1.186e-02
  model.layers.3.self_attn.v_proj                           161      0     1.263e-02    2.311e-03    1.284e-02
  model.layers.30.linear_attn.in_proj_qkv                   161      0     1.189e-02    2.230e-03    1.210e-02
  model.layers.30.linear_attn.in_proj_z                     161      0     1.349e-02    1.886e-03    1.362e-02
  model.layers.30.linear_attn.out_proj                      161      0     1.622e-02    2.412e-03    1.640e-02
  model.layers.30.mlp.down_proj                             161      0     1.285e-02    2.561e-03    1.310e-02
  model.layers.30.mlp.gate_proj                             161      0     1.171e-02    3.188e-03    1.214e-02
  model.layers.30.mlp.up_proj                               161      0     1.067e-02    3.521e-03    1.123e-02
  model.layers.31.mlp.down_proj                             161      0     1.320e-02    2.596e-03    1.345e-02
  model.layers.31.mlp.gate_proj                             161      0     1.148e-02    3.154e-03    1.191e-02
  model.layers.31.mlp.up_proj                               161      0     1.090e-02    3.518e-03    1.146e-02
  model.layers.31.self_attn.k_proj                          161      0     1.207e-02    2.519e-03    1.233e-02
  model.layers.31.self_attn.o_proj                          161      0     2.214e-02    3.029e-03    2.235e-02
  model.layers.31.self_attn.q_proj                          161      0     1.015e-02    2.074e-03    1.036e-02
  model.layers.31.self_attn.v_proj                          161      0     9.912e-03    2.191e-03    1.014e-02
  model.layers.32.linear_attn.in_proj_qkv                   161      0     1.282e-02    2.287e-03    1.303e-02
  model.layers.32.linear_attn.in_proj_z                     161      0     1.411e-02    1.985e-03    1.425e-02
  model.layers.32.linear_attn.out_proj                      161      0     1.567e-02    2.303e-03    1.585e-02
  model.layers.32.mlp.down_proj                             161      0     1.301e-02    2.618e-03    1.327e-02
  model.layers.32.mlp.gate_proj                             161      0     1.165e-02    3.107e-03    1.206e-02
  model.layers.32.mlp.up_proj                               161      0     1.108e-02    3.522e-03    1.162e-02
  model.layers.33.linear_attn.in_proj_qkv                   161      0     1.259e-02    2.298e-03    1.280e-02
  model.layers.33.linear_attn.in_proj_z                     161      0     1.386e-02    1.988e-03    1.400e-02
  model.layers.33.linear_attn.out_proj                      161      0     1.745e-02    2.367e-03    1.761e-02
  model.layers.33.mlp.down_proj                             161      0     1.342e-02    2.602e-03    1.367e-02
  model.layers.33.mlp.gate_proj                             161      0     1.193e-02    3.114e-03    1.232e-02
  model.layers.33.mlp.up_proj                               161      0     1.124e-02    3.542e-03    1.178e-02
  model.layers.34.linear_attn.in_proj_qkv                   161      0     1.252e-02    2.278e-03    1.273e-02
  model.layers.34.linear_attn.in_proj_z                     161      0     1.343e-02    1.914e-03    1.357e-02
  model.layers.34.linear_attn.out_proj                      161      0     1.609e-02    2.363e-03    1.626e-02
  model.layers.34.mlp.down_proj                             161      0     1.377e-02    2.610e-03    1.401e-02
  model.layers.34.mlp.gate_proj                             161      0     1.207e-02    3.070e-03    1.245e-02
  model.layers.34.mlp.up_proj                               161      0     1.115e-02    3.505e-03    1.169e-02
  model.layers.35.mlp.down_proj                             161      0     1.414e-02    2.562e-03    1.437e-02
  model.layers.35.mlp.gate_proj                             161      0     1.066e-02    3.068e-03    1.109e-02
  model.layers.35.mlp.up_proj                               161      0     1.138e-02    3.583e-03    1.193e-02
  model.layers.35.self_attn.k_proj                          161      0     1.177e-02    2.500e-03    1.203e-02
  model.layers.35.self_attn.o_proj                          161      0     1.997e-02    2.809e-03    2.017e-02
  model.layers.35.self_attn.q_proj                          161      0     1.072e-02    2.117e-03    1.092e-02
  model.layers.35.self_attn.v_proj                          161      0     1.071e-02    2.262e-03    1.094e-02
  model.layers.36.linear_attn.in_proj_qkv                   161      0     1.295e-02    2.276e-03    1.314e-02
  model.layers.36.linear_attn.in_proj_z                     161      0     1.212e-02    1.746e-03    1.224e-02
  model.layers.36.linear_attn.out_proj                      161      0     1.667e-02    2.261e-03    1.682e-02
  model.layers.36.mlp.down_proj                             161      0     1.511e-02    2.627e-03    1.534e-02
  model.layers.36.mlp.gate_proj                             161      0     1.026e-02    3.145e-03    1.074e-02
  model.layers.36.mlp.up_proj                               161      0     1.154e-02    3.603e-03    1.209e-02
  model.layers.37.linear_attn.in_proj_qkv                   161      0     1.223e-02    2.242e-03    1.243e-02
  model.layers.37.linear_attn.in_proj_z                     161      0     1.058e-02    1.597e-03    1.071e-02
  model.layers.37.linear_attn.out_proj                      161      0     1.838e-02    2.199e-03    1.851e-02
  model.layers.37.mlp.down_proj                             161      0     1.541e-02    2.734e-03    1.565e-02
  model.layers.37.mlp.gate_proj                             161      0     1.055e-02    3.235e-03    1.104e-02
  model.layers.37.mlp.up_proj                               161      0     1.172e-02    3.628e-03    1.227e-02
  model.layers.38.linear_attn.in_proj_qkv                   161      0     1.201e-02    2.319e-03    1.223e-02
  model.layers.38.linear_attn.in_proj_z                     161      0     1.100e-02    1.630e-03    1.112e-02
  model.layers.38.linear_attn.out_proj                      161      0     1.656e-02    2.310e-03    1.672e-02
  model.layers.38.mlp.down_proj                             161      0     1.595e-02    2.700e-03    1.617e-02
  model.layers.38.mlp.gate_proj                             161      0     1.069e-02    3.065e-03    1.112e-02
  model.layers.38.mlp.up_proj                               161      0     1.171e-02    3.642e-03    1.226e-02
  model.layers.39.mlp.down_proj                             161      0     1.441e-02    2.655e-03    1.465e-02
  model.layers.39.mlp.gate_proj                             161      0     9.965e-03    3.158e-03    1.045e-02
  model.layers.39.mlp.up_proj                               161      0     1.180e-02    3.648e-03    1.235e-02
  model.layers.39.self_attn.k_proj                          161      0     1.179e-02    2.470e-03    1.205e-02
  model.layers.39.self_attn.o_proj                          161      0     2.077e-02    2.841e-03    2.097e-02
  model.layers.39.self_attn.q_proj                          161      0     1.045e-02    2.093e-03    1.066e-02
  model.layers.39.self_attn.v_proj                          161      0     1.090e-02    2.329e-03    1.114e-02
  model.layers.4.linear_attn.in_proj_qkv                    161      0     1.192e-02    2.310e-03    1.214e-02
  model.layers.4.linear_attn.in_proj_z                      161      0     1.247e-02    1.774e-03    1.260e-02
  model.layers.4.linear_attn.out_proj                       161      0     1.788e-02    2.410e-03    1.804e-02
  model.layers.4.mlp.down_proj                              161      0     1.444e-02    2.610e-03    1.468e-02
  model.layers.4.mlp.gate_proj                              161      0     1.015e-02    3.511e-03    1.074e-02
  model.layers.4.mlp.up_proj                                161      0     1.123e-02    3.616e-03    1.180e-02
  model.layers.40.linear_attn.in_proj_qkv                   161      0     1.226e-02    2.281e-03    1.247e-02
  model.layers.40.linear_attn.in_proj_z                     161      0     1.172e-02    1.686e-03    1.185e-02
  model.layers.40.linear_attn.out_proj                      161      0     1.611e-02    2.266e-03    1.626e-02
  model.layers.40.mlp.down_proj                             161      0     1.485e-02    2.675e-03    1.508e-02
  model.layers.40.mlp.gate_proj                             161      0     1.008e-02    3.132e-03    1.056e-02
  model.layers.40.mlp.up_proj                               161      0     1.164e-02    3.641e-03    1.220e-02
  model.layers.41.linear_attn.in_proj_qkv                   161      0     1.175e-02    2.257e-03    1.197e-02
  model.layers.41.linear_attn.in_proj_z                     161      0     1.225e-02    1.732e-03    1.238e-02
  model.layers.41.linear_attn.out_proj                      161      0     1.646e-02    2.257e-03    1.662e-02
  model.layers.41.mlp.down_proj                             161      0     1.475e-02    2.702e-03    1.500e-02
  model.layers.41.mlp.gate_proj                             161      0     1.085e-02    3.212e-03    1.131e-02
  model.layers.41.mlp.up_proj                               161      0     1.166e-02    3.633e-03    1.222e-02
  model.layers.42.linear_attn.in_proj_qkv                   161      0     1.163e-02    2.277e-03    1.185e-02
  model.layers.42.linear_attn.in_proj_z                     161      0     1.240e-02    1.748e-03    1.253e-02
  model.layers.42.linear_attn.out_proj                      161      0     1.574e-02    2.284e-03    1.591e-02
  model.layers.42.mlp.down_proj                             161      0     1.476e-02    2.644e-03    1.499e-02
  model.layers.42.mlp.gate_proj                             161      0     1.173e-02    3.240e-03    1.217e-02
  model.layers.42.mlp.up_proj                               161      0     1.143e-02    3.584e-03    1.198e-02
  model.layers.43.mlp.down_proj                             161      0     1.381e-02    2.670e-03    1.407e-02
  model.layers.43.mlp.gate_proj                             161      0     1.085e-02    3.218e-03    1.132e-02
  model.layers.43.mlp.up_proj                               161      0     1.154e-02    3.618e-03    1.210e-02
  model.layers.43.self_attn.k_proj                          161      0     1.253e-02    2.452e-03    1.276e-02
  model.layers.43.self_attn.o_proj                          161      0     2.163e-02    3.125e-03    2.186e-02
  model.layers.43.self_attn.q_proj                          161      0     1.062e-02    2.091e-03    1.082e-02
  model.layers.43.self_attn.v_proj                          161      0     1.054e-02    2.252e-03    1.078e-02
  model.layers.44.linear_attn.in_proj_qkv                   161      0     1.206e-02    2.230e-03    1.226e-02
  model.layers.44.linear_attn.in_proj_z                     161      0     1.228e-02    1.698e-03    1.240e-02
  model.layers.44.linear_attn.out_proj                      161      0     1.628e-02    2.298e-03    1.644e-02
  model.layers.44.mlp.down_proj                             161      0     1.414e-02    2.712e-03    1.440e-02
  model.layers.44.mlp.gate_proj                             161      0     1.083e-02    3.259e-03    1.131e-02
  model.layers.44.mlp.up_proj                               161      0     1.171e-02    3.618e-03    1.225e-02
  model.layers.45.linear_attn.in_proj_qkv                   161      0     1.209e-02    2.277e-03    1.230e-02
  model.layers.45.linear_attn.in_proj_z                     161      0     1.265e-02    1.759e-03    1.277e-02
  model.layers.45.linear_attn.out_proj                      161      0     1.814e-02    2.430e-03    1.830e-02
  model.layers.45.mlp.down_proj                             161      0     1.437e-02    2.740e-03    1.463e-02
  model.layers.45.mlp.gate_proj                             161      0     1.082e-02    3.289e-03    1.131e-02
  model.layers.45.mlp.up_proj                               161      0     1.157e-02    3.623e-03    1.212e-02
  model.layers.46.linear_attn.in_proj_qkv                   161      0     1.198e-02    2.245e-03    1.219e-02
  model.layers.46.linear_attn.in_proj_z                     161      0     1.186e-02    1.650e-03    1.197e-02
  model.layers.46.linear_attn.out_proj                      161      0     1.868e-02    2.358e-03    1.883e-02
  model.layers.46.mlp.down_proj                             161      0     1.426e-02    2.688e-03    1.451e-02
  model.layers.46.mlp.gate_proj                             161      0     1.112e-02    3.310e-03    1.161e-02
  model.layers.46.mlp.up_proj                               161      0     1.151e-02    3.600e-03    1.206e-02
  model.layers.47.mlp.down_proj                             161      0     1.426e-02    2.639e-03    1.450e-02
  model.layers.47.mlp.gate_proj                             161      0     1.057e-02    3.335e-03    1.108e-02
  model.layers.47.mlp.up_proj                               161      0     1.177e-02    3.621e-03    1.232e-02
  model.layers.47.self_attn.k_proj                          161      0     1.248e-02    2.462e-03    1.272e-02
  model.layers.47.self_attn.o_proj                          161      0     1.954e-02    2.555e-03    1.971e-02
  model.layers.47.self_attn.q_proj                          161      0     9.680e-03    2.056e-03    9.896e-03
  model.layers.47.self_attn.v_proj                          161      0     1.068e-02    2.234e-03    1.091e-02
  model.layers.48.linear_attn.in_proj_qkv                   161      0     1.274e-02    2.267e-03    1.294e-02
  model.layers.48.linear_attn.in_proj_z                     161      0     1.210e-02    1.652e-03    1.221e-02
  model.layers.48.linear_attn.out_proj                      161      0     1.720e-02    2.416e-03    1.737e-02
  model.layers.48.mlp.down_proj                             161      0     1.479e-02    2.684e-03    1.503e-02
  model.layers.48.mlp.gate_proj                             161      0     1.036e-02    3.332e-03    1.088e-02
  model.layers.48.mlp.up_proj                               161      0     1.170e-02    3.606e-03    1.224e-02
  model.layers.49.linear_attn.in_proj_qkv                   161      0     1.211e-02    2.277e-03    1.232e-02
  model.layers.49.linear_attn.in_proj_z                     161      0     1.156e-02    1.608e-03    1.167e-02
  model.layers.49.linear_attn.out_proj                      161      0     1.910e-02    2.175e-03    1.923e-02
  model.layers.49.mlp.down_proj                             161      0     1.567e-02    2.697e-03    1.590e-02
  model.layers.49.mlp.gate_proj                             161      0     1.007e-02    3.367e-03    1.062e-02
  model.layers.49.mlp.up_proj                               161      0     1.150e-02    3.563e-03    1.204e-02
  model.layers.5.linear_attn.in_proj_qkv                    161      0     1.245e-02    2.270e-03    1.266e-02
  model.layers.5.linear_attn.in_proj_z                      161      0     1.391e-02    1.910e-03    1.404e-02
  model.layers.5.linear_attn.out_proj                       161      0     1.772e-02    2.209e-03    1.785e-02
  model.layers.5.mlp.down_proj                              161      0     1.529e-02    2.519e-03    1.550e-02
  model.layers.5.mlp.gate_proj                              161      0     1.020e-02    3.534e-03    1.079e-02
  model.layers.5.mlp.up_proj                                161      0     1.128e-02    3.622e-03    1.185e-02
  model.layers.50.linear_attn.in_proj_qkv                   161      0     1.169e-02    2.255e-03    1.190e-02
  model.layers.50.linear_attn.in_proj_z                     161      0     1.090e-02    1.493e-03    1.100e-02
  model.layers.50.linear_attn.out_proj                      161      0     1.824e-02    2.348e-03    1.839e-02
  model.layers.50.mlp.down_proj                             161      0     1.517e-02    2.666e-03    1.541e-02
  model.layers.50.mlp.gate_proj                             161      0     9.861e-03    3.280e-03    1.039e-02
  model.layers.50.mlp.up_proj                               161      0     1.087e-02    3.439e-03    1.140e-02
  model.layers.51.mlp.down_proj                             161      0     1.524e-02    2.656e-03    1.548e-02
  model.layers.51.mlp.gate_proj                             161      0     8.699e-03    3.392e-03    9.339e-03
  model.layers.51.mlp.up_proj                               161      0     1.111e-02    3.552e-03    1.166e-02
  model.layers.51.self_attn.k_proj                          161      0     1.103e-02    2.402e-03    1.129e-02
  model.layers.51.self_attn.o_proj                          161      0     1.919e-02    2.299e-03    1.933e-02
  model.layers.51.self_attn.q_proj                          161      0     9.135e-03    2.024e-03    9.356e-03
  model.layers.51.self_attn.v_proj                          161      0     1.024e-02    2.274e-03    1.049e-02
  model.layers.52.linear_attn.in_proj_qkv                   161      0     9.642e-03    2.110e-03    9.868e-03
  model.layers.52.linear_attn.in_proj_z                     161      0     8.558e-03    1.286e-03    8.653e-03
  model.layers.52.linear_attn.out_proj                      161      0     1.877e-02    2.303e-03    1.891e-02
  model.layers.52.mlp.down_proj                             161      0     1.784e-02    2.685e-03    1.804e-02
  model.layers.52.mlp.gate_proj                             161      0     8.139e-03    3.398e-03    8.821e-03
  model.layers.52.mlp.up_proj                               161      0     1.105e-02    3.564e-03    1.161e-02
  model.layers.53.linear_attn.in_proj_qkv                   161      0     9.505e-03    2.128e-03    9.740e-03
  model.layers.53.linear_attn.in_proj_z                     161      0     8.688e-03    1.313e-03    8.786e-03
  model.layers.53.linear_attn.out_proj                      161      0     1.699e-02    2.275e-03    1.714e-02
  model.layers.53.mlp.down_proj                             161      0     1.846e-02    2.768e-03    1.867e-02
  model.layers.53.mlp.gate_proj                             161      0     7.902e-03    3.398e-03    8.601e-03
  model.layers.53.mlp.up_proj                               161      0     1.096e-02    3.568e-03    1.153e-02
  model.layers.54.linear_attn.in_proj_qkv                   161      0     9.920e-03    2.231e-03    1.017e-02
  model.layers.54.linear_attn.in_proj_z                     161      0     8.309e-03    1.278e-03    8.406e-03
  model.layers.54.linear_attn.out_proj                      161      0     1.749e-02    2.127e-03    1.762e-02
  model.layers.54.mlp.down_proj                             161      0     1.815e-02    2.477e-03    1.832e-02
  model.layers.54.mlp.gate_proj                             161      0     7.284e-03    3.349e-03    8.016e-03
  model.layers.54.mlp.up_proj                               161      0     9.161e-03    3.418e-03    9.777e-03
  model.layers.55.mlp.down_proj                             161      0     1.696e-02    2.620e-03    1.716e-02
  model.layers.55.mlp.gate_proj                             161      0     7.234e-03    3.384e-03    7.986e-03
  model.layers.55.mlp.up_proj                               161      0     1.132e-02    3.603e-03    1.188e-02
  model.layers.55.self_attn.k_proj                          161      0     1.085e-02    2.400e-03    1.112e-02
  model.layers.55.self_attn.o_proj                          161      0     1.731e-02    2.199e-03    1.745e-02
  model.layers.55.self_attn.q_proj                          161      0     8.915e-03    2.106e-03    9.160e-03
  model.layers.55.self_attn.v_proj                          161      0     9.482e-03    2.219e-03    9.739e-03
  model.layers.56.linear_attn.in_proj_qkv                   161      0     9.010e-03    2.090e-03    9.249e-03
  model.layers.56.linear_attn.in_proj_z                     161      0     8.724e-03    1.331e-03    8.825e-03
  model.layers.56.linear_attn.out_proj                      161      0     1.741e-02    2.098e-03    1.754e-02
  model.layers.56.mlp.down_proj                             161      0     1.968e-02    2.530e-03    1.984e-02
  model.layers.56.mlp.gate_proj                             161      0     7.328e-03    3.347e-03    8.057e-03
  model.layers.56.mlp.up_proj                               161      0     1.133e-02    3.595e-03    1.188e-02
  model.layers.57.linear_attn.in_proj_qkv                   161      0     8.646e-03    2.053e-03    8.886e-03
  model.layers.57.linear_attn.in_proj_z                     161      0     8.532e-03    1.300e-03    8.631e-03
  model.layers.57.linear_attn.out_proj                      161      0     1.595e-02    2.510e-03    1.615e-02
  model.layers.57.mlp.down_proj                             161      0     1.634e-02    2.796e-03    1.657e-02
  model.layers.57.mlp.gate_proj                             161      0     7.144e-03    3.321e-03    7.878e-03
  model.layers.57.mlp.up_proj                               161      0     1.116e-02    3.600e-03    1.173e-02
  model.layers.58.linear_attn.in_proj_qkv                   161      0     9.203e-03    2.131e-03    9.447e-03
  model.layers.58.linear_attn.in_proj_z                     161      0     8.789e-03    1.313e-03    8.885e-03
  model.layers.58.linear_attn.out_proj                      161      0     1.895e-02    2.411e-03    1.910e-02
  model.layers.58.mlp.down_proj                             161      0     1.666e-02    2.607e-03    1.686e-02
  model.layers.58.mlp.gate_proj                             161      0     7.020e-03    3.313e-03    7.763e-03
  model.layers.58.mlp.up_proj                               161      0     1.044e-02    3.547e-03    1.102e-02
  model.layers.59.mlp.down_proj                             161      0     1.667e-02    2.502e-03    1.685e-02
  model.layers.59.mlp.gate_proj                             161      0     6.604e-03    3.306e-03    7.388e-03
  model.layers.59.mlp.up_proj                               161      0     9.269e-03    3.450e-03    9.890e-03
  model.layers.59.self_attn.k_proj                          161      0     1.016e-02    2.336e-03    1.043e-02
  model.layers.59.self_attn.o_proj                          161      0     1.802e-02    2.288e-03    1.816e-02
  model.layers.59.self_attn.q_proj                          161      0     8.663e-03    2.118e-03    8.920e-03
  model.layers.59.self_attn.v_proj                          161      0     9.077e-03    2.160e-03    9.329e-03
  model.layers.6.linear_attn.in_proj_qkv                    161      0     1.263e-02    2.269e-03    1.283e-02
  model.layers.6.linear_attn.in_proj_z                      161      0     1.427e-02    1.989e-03    1.441e-02
  model.layers.6.linear_attn.out_proj                       161      0     1.682e-02    2.292e-03    1.697e-02
  model.layers.6.mlp.down_proj                              161      0     1.358e-02    2.400e-03    1.379e-02
  model.layers.6.mlp.gate_proj                              161      0     1.045e-02    3.493e-03    1.102e-02
  model.layers.6.mlp.up_proj                                161      0     1.113e-02    3.563e-03    1.168e-02
  model.layers.60.linear_attn.in_proj_qkv                   161      0     7.947e-03    1.973e-03    8.189e-03
  model.layers.60.linear_attn.in_proj_z                     161      0     7.538e-03    1.195e-03    7.633e-03
  model.layers.60.linear_attn.out_proj                      161      0     1.696e-02    2.504e-03    1.714e-02
  model.layers.60.mlp.down_proj                             161      0     1.583e-02    2.704e-03    1.606e-02
  model.layers.60.mlp.gate_proj                             161      0     6.834e-03    3.243e-03    7.564e-03
  model.layers.60.mlp.up_proj                               161      0     9.170e-03    3.388e-03    9.776e-03
  model.layers.61.linear_attn.in_proj_qkv                   161      0     8.878e-03    2.088e-03    9.120e-03
  model.layers.61.linear_attn.in_proj_z                     161      0     7.963e-03    1.246e-03    8.060e-03
  model.layers.61.linear_attn.out_proj                      161      0     1.669e-02    2.348e-03    1.686e-02
  model.layers.61.mlp.down_proj                             161      0     1.498e-02    2.664e-03    1.521e-02
  model.layers.61.mlp.gate_proj                             161      0     6.562e-03    3.195e-03    7.300e-03
  model.layers.61.mlp.up_proj                               161      0     8.694e-03    3.315e-03    9.304e-03
  model.layers.62.linear_attn.in_proj_qkv                   161      0     8.149e-03    1.974e-03    8.384e-03
  model.layers.62.linear_attn.in_proj_z                     161      0     7.410e-03    1.221e-03    7.508e-03
  model.layers.62.linear_attn.out_proj                      161      0     1.516e-02    2.354e-03    1.535e-02
  model.layers.62.mlp.down_proj                             161      0     1.496e-02    2.533e-03    1.518e-02
  model.layers.62.mlp.gate_proj                             161      0     6.266e-03    3.134e-03    7.006e-03
  model.layers.62.mlp.up_proj                               161      0     6.854e-03    3.145e-03    7.542e-03
  model.layers.63.mlp.down_proj                             161      0     1.191e-02    2.011e-03    1.208e-02
  model.layers.63.mlp.gate_proj                             161      0     5.660e-03    3.068e-03    6.439e-03
  model.layers.63.mlp.up_proj                               161      0     4.709e-03    3.050e-03    5.610e-03
  model.layers.63.self_attn.k_proj                          161      0     1.072e-02    2.414e-03    1.098e-02
  model.layers.63.self_attn.o_proj                          161      0     1.293e-02    2.012e-03    1.308e-02
  model.layers.63.self_attn.q_proj                          161      0     7.013e-03    1.939e-03    7.277e-03
  model.layers.63.self_attn.v_proj                          161      0     4.703e-03    1.727e-03    5.007e-03
  model.layers.7.mlp.down_proj                              161      0     1.425e-02    2.578e-03    1.448e-02
  model.layers.7.mlp.gate_proj                              161      0     1.041e-02    3.566e-03    1.100e-02
  model.layers.7.mlp.up_proj                                161      0     1.135e-02    3.626e-03    1.191e-02
  model.layers.7.self_attn.k_proj                           161      0     1.383e-02    2.714e-03    1.410e-02
  model.layers.7.self_attn.o_proj                           161      0     1.329e-02    2.267e-03    1.347e-02
  model.layers.7.self_attn.q_proj                           161      0     1.250e-02    2.079e-03    1.267e-02
  model.layers.7.self_attn.v_proj                           161      0     1.102e-02    2.332e-03    1.126e-02
  model.layers.8.linear_attn.in_proj_qkv                    161      0     1.243e-02    2.248e-03    1.263e-02
  model.layers.8.linear_attn.in_proj_z                      161      0     1.308e-02    1.811e-03    1.320e-02
  model.layers.8.linear_attn.out_proj                       161      0     1.617e-02    2.281e-03    1.632e-02
  model.layers.8.mlp.down_proj                              161      0     1.507e-02    2.557e-03    1.528e-02
  model.layers.8.mlp.gate_proj                              161      0     1.051e-02    3.563e-03    1.109e-02
  model.layers.8.mlp.up_proj                                161      0     1.129e-02    3.619e-03    1.186e-02
  model.layers.9.linear_attn.in_proj_qkv                    161      0     1.309e-02    2.233e-03    1.328e-02
  model.layers.9.linear_attn.in_proj_z                      161      0     1.361e-02    1.895e-03    1.374e-02
  model.layers.9.linear_attn.out_proj                       161      0     1.602e-02    2.267e-03    1.618e-02
  model.layers.9.mlp.down_proj                              161      0     1.345e-02    2.475e-03    1.368e-02
  model.layers.9.mlp.gate_proj                              161      0     1.058e-02    3.577e-03    1.117e-02
  model.layers.9.mlp.up_proj                                161      0     1.119e-02    3.610e-03    1.176e-02
[i8] per-arm census (int8 kernel calls): {'REF': 0, 'REF2': 0, 'STD': 0, 'SIMROW': 0, 'INT8': 64000}
[i8] STD control: fused-reconstruct override consulted 64000 times

ARM RECEIPTS -- what each arm actually executed, from the engine's
own per-call census.  fused/standalone is the reconstruct variant;
acc16/acc32 is the cuBLAS compute type leg B asked hgemm for.
  arm       legB calls          rows     fused  standalone     acc16     acc32   fp32pin
  REF            64000     131072000     64000           0     64000         0         0
  REF2           64000     131072000     64000           0     64000         0         0
  STD            64000     131072000         0       64000     64000         0         0
  SIMROW         64000     131072000         0       64000     64000         0         0
  INT8           64000     131072000         0       64000     64000         0         0
  all arms match their expected leg/accumulator signature
  prefill positions aligned: [8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8] per prompt, identical across all 5 arms
[i8] null control: REF2 vs REF max KL is exactly 0

THE TAIL over PREFILL positions only -- one per prefill chunk, so
every arm scores the SAME context. Decode positions are reported
separately and are NOT a numerics measure: once a token flips, the
arms are reading different contexts. The control proves it -- pooled
over decode, STD (which contains no int8 at all) has the same max KL
as INT8, so a pooled tail cannot attribute anything to the kernel.

arm            n        mean         p50         p95         p99         MAX     argmax flip
REF          160   0.000e+00   0.000e+00   0.000e+00   0.000e+00   0.000e+00 0/160 =   0.00%
REF2         160   0.000e+00   0.000e+00   0.000e+00   0.000e+00   0.000e+00 0/160 =   0.00%
STD          160   1.571e-02   3.134e-04   8.107e-03   7.665e-01   1.167e+00 4/160 =   2.50%
SIMROW       160   1.078e-02   7.276e-04   1.418e-02   2.130e-01   1.044e+00 6/160 =   3.75%
INT8         160   7.370e-02   9.407e-04   1.852e-02   2.040e+00   8.940e+00 5/160 =   3.12%

Tail attribution -- the 10 worst INT8 prefill positions, with the
control's KL at the SAME position. A tail that is positional shows
the control elevated too; a tail that is the kernel's does not.
   prompt  chunk      INT8 KL       STD KL     SIMROW KL  ratio INT8/STD
       16      6    8.940e+00    2.786e-02     5.223e-02          320.88
        0      7    2.040e+00    1.167e+00     2.130e-01            1.75
        0      1    3.977e-01    7.665e-01     1.044e+00            0.52
        2      7    5.191e-02    1.035e-03     4.090e-03           50.15
        2      3    2.703e-02    3.618e-01     1.030e-03            0.07
        0      2    2.209e-02    2.338e-03     8.927e-03            9.45
       18      7    2.161e-02    1.509e-03     5.880e-03           14.32
       14      3    1.852e-02    1.122e-03     4.090e-03           16.50
        3      4    1.831e-02    2.425e-04     2.784e-02           75.51
       17      5    1.685e-02    9.180e-03     7.866e-03            1.84
  Spearman rank correlation STD vs INT8 across all 160 prefill positions: 0.817

Decode positions (divergence-contaminated, shown for completeness):
arm            n        mean         MAX    greedy agree
REF          160   0.000e+00   0.000e+00 160/160 =  100.0%
REF2         160   0.000e+00   0.000e+00 160/160 =  100.0%
STD          160   4.864e-01   2.361e+01 156/160 =   97.5%
SIMROW       160   1.850e+00   3.466e+01 147/160 =   91.9%
INT8         160   2.621e+00   3.244e+01 137/160 =   85.6%

SAMPLER-RELEVANT TAIL @ canonical (T=1.0, top_p=0.95, top_k=20), prefill positions only.
TV is the headline: under a drafter that tracks the target, speculative
decoding's expected accept rate is sum_t min(p_ref, p_arm), so TV IS the
expected accept-rate loss.
arm            n  setOverlap    TV mean     TV p95     TV max  truncKL p95  bndryChurn  drawDiff
REF          160     100.00%  0.000e+00  0.000e+00  0.000e+00    0.000e+00 0/160     0.00%
REF2         160     100.00%  0.000e+00  0.000e+00  0.000e+00    0.000e+00 0/160     0.00%
STD          160      98.04%  1.587e-02  4.462e-02  4.939e-01          inf 26/160     3.52%
SIMROW       160      97.92%  2.079e-02  7.235e-02  4.378e-01          inf 43/160     3.98%
INT8         160      98.01%  2.699e-02  6.755e-02  9.229e-01          inf 40/160     5.62%

SAMPLER-RELEVANT TAIL @ owner's k~40 (T=1.0, top_p=1.0, top_k=40), prefill positions only.
TV is the headline: under a drafter that tracks the target, speculative
decoding's expected accept rate is sum_t min(p_ref, p_arm), so TV IS the
expected accept-rate loss.
arm            n  setOverlap    TV mean     TV p95     TV max  truncKL p95  bndryChurn  drawDiff
REF          160     100.00%  0.000e+00  0.000e+00  0.000e+00    0.000e+00 0/160     0.00%
REF2         160     100.00%  0.000e+00  0.000e+00  0.000e+00    0.000e+00 0/160     0.00%
STD          160      97.42%  1.573e-02  3.820e-02  4.806e-01          inf 86/160     4.61%
SIMROW       160      96.40%  2.008e-02  6.820e-02  4.394e-01          inf 110/160     7.19%
INT8         160      95.65%  2.644e-02  6.605e-02  9.047e-01          inf 121/160     7.97%

FULL-VOCAB numbers below, kept and labelled: they are the right
instrument for GREEDY or beam decode, where every rank matters.

arm         first-prefill-position KL       median          max
REF                         0.000e+00    0.000e+00    0.000e+00
REF2                        0.000e+00    0.000e+00    0.000e+00
STD                         3.479e-04    3.551e-05    3.823e-03
SIMROW                      5.748e-04    1.636e-04    2.633e-03
INT8                        6.910e-04    1.047e-04    2.922e-03

LAYER TRACE, prompt 16 -- rel rms against REF of the residual
stream leaving each decoder layer, at the token the KL is scored
at.  A STEP at one layer is a defect with an address; smooth
multiplicative growth with no step is the model amplifying.
  per-chunk full-vocab KL vs REF on this prompt:
  chunk           REF        REF2         STD      SIMROW        INT8
      0     0.000e+00   0.000e+00   5.943e-05   6.153e-05   7.360e-05
      1     0.000e+00   0.000e+00   5.318e-04   2.382e-04   1.031e-03
      2     0.000e+00   0.000e+00   5.195e-05   4.302e-05   9.914e-04
      3     0.000e+00   0.000e+00   1.018e-03   7.363e-03   2.825e-03
      4     0.000e+00   0.000e+00   5.652e-04   5.537e-03   2.387e-03
      5     0.000e+00   0.000e+00   1.371e-06   1.420e-06   4.626e-07
      6     0.000e+00   0.000e+00   2.786e-02   5.223e-02   8.940e+00
      7     0.000e+00   0.000e+00   8.139e-04   4.692e-03   1.156e-02
  worst chunk for INT8: c6

  chunk 0: rel rms after layer L, vs REF
  layer        REF2         STD      SIMROW        INT8     ratio
      0   0.000e+00   8.725e-03   2.074e-02   2.337e-02       nan
      1   0.000e+00   1.550e-02   2.324e-02   2.548e-02       nan
      2   0.000e+00   2.111e-02   2.279e-02   1.879e-02       nan
      3   0.000e+00   1.177e-02   2.961e-02   3.201e-02       nan
      4   0.000e+00   1.948e-02   3.022e-02   3.459e-02       nan
      5   0.000e+00   1.927e-02   3.104e-02   4.179e-02       nan
      6   0.000e+00   1.124e-02   2.680e-02   2.996e-02       nan
      7   0.000e+00   1.471e-02   2.984e-02   3.399e-02       nan
      8   0.000e+00   1.191e-02   2.831e-02   3.565e-02       nan
      9   0.000e+00   8.817e-03   2.618e-02   3.329e-02       nan
     10   0.000e+00   1.240e-02   2.758e-02   3.644e-02       nan
     11   0.000e+00   1.244e-02   2.696e-02   3.360e-02       nan
     12   0.000e+00   1.389e-02   2.860e-02   3.648e-02       nan
     13   0.000e+00   1.545e-02   3.175e-02   3.872e-02       nan
     14   0.000e+00   1.429e-02   2.937e-02   3.515e-02       nan
     15   0.000e+00   1.653e-02   3.695e-02   4.196e-02       nan
     16   0.000e+00   1.591e-02   3.397e-02   3.860e-02       nan
     17   0.000e+00   1.285e-02   2.895e-02   3.370e-02       nan
     18   0.000e+00   1.021e-02   2.022e-02   2.337e-02       nan
     19   0.000e+00   1.169e-02   2.668e-02   3.162e-02       nan
     20   0.000e+00   1.146e-02   2.786e-02   3.235e-02       nan
     21   0.000e+00   1.313e-02   2.907e-02   3.292e-02       nan
     22   0.000e+00   1.245e-02   3.008e-02   3.104e-02       nan
     23   0.000e+00   1.340e-02   3.316e-02   3.699e-02       nan
     24   0.000e+00   1.345e-02   3.518e-02   3.707e-02       nan
     25   0.000e+00   1.422e-02   3.119e-02   3.778e-02       nan
     26   0.000e+00   1.261e-02   2.874e-02   3.076e-02       nan
     27   0.000e+00   1.532e-02   3.225e-02   3.573e-02       nan
     28   0.000e+00   1.728e-02   3.528e-02   4.105e-02       nan
     29   0.000e+00   1.550e-02   3.454e-02   3.718e-02       nan
     30   0.000e+00   1.527e-02   3.426e-02   3.807e-02       nan
     31   0.000e+00   1.692e-02   3.652e-02   4.158e-02       nan
     32   0.000e+00   2.046e-02   3.899e-02   4.314e-02       nan
     33   0.000e+00   2.099e-02   4.155e-02   4.528e-02       nan
     34   0.000e+00   2.034e-02   4.037e-02   4.299e-02       nan
     35   0.000e+00   1.979e-02   3.741e-02   4.102e-02       nan
     36   0.000e+00   2.114e-02   4.068e-02   4.601e-02       nan
     37   0.000e+00   2.253e-02   4.596e-02   4.938e-02       nan
     38   0.000e+00   2.346e-02   4.494e-02   4.740e-02       nan
     39   0.000e+00   2.172e-02   3.818e-02   4.250e-02       nan
     40   0.000e+00   2.436e-02   4.511e-02   4.853e-02       nan
     41   0.000e+00   2.566e-02   4.570e-02   4.809e-02       nan
     42   0.000e+00   2.017e-02   4.118e-02   4.653e-02       nan
     43   0.000e+00   2.789e-02   5.656e-02   6.366e-02       nan
     44   0.000e+00   2.393e-02   5.347e-02   5.940e-02       nan
     45   0.000e+00   2.541e-02   5.627e-02   6.064e-02       nan
     46   0.000e+00   2.645e-02   4.910e-02   5.793e-02       nan
     47   0.000e+00   2.372e-02   5.197e-02   5.316e-02       nan
     48   0.000e+00   2.221e-02   4.804e-02   5.554e-02       nan
     49   0.000e+00   1.979e-02   3.739e-02   4.374e-02       nan
     50   0.000e+00   1.901e-02   3.287e-02   3.559e-02       nan
     51   0.000e+00   2.165e-02   3.778e-02   4.158e-02       nan
     52   0.000e+00   1.781e-02   3.866e-02   4.152e-02       nan
     53   0.000e+00   2.185e-02   3.634e-02   4.430e-02       nan
     54   0.000e+00   1.736e-02   3.429e-02   3.634e-02       nan
     55   0.000e+00   1.990e-02   3.611e-02   3.917e-02       nan
     56   0.000e+00   2.071e-02   3.236e-02   4.073e-02       nan
     57   0.000e+00   2.325e-02   4.010e-02   4.687e-02       nan
     58   0.000e+00   2.168e-02   3.199e-02   3.911e-02       nan
     59   0.000e+00   1.672e-02   2.852e-02   3.733e-02       nan
     60   0.000e+00   2.207e-02   3.134e-02   4.511e-02       nan
     61   0.000e+00   1.951e-02   2.816e-02   3.396e-02       nan
     62   0.000e+00   2.914e-02   3.171e-02   4.600e-02       nan
     63   0.000e+00   2.906e-02   3.771e-02   3.735e-02       nan

  chunk 5: rel rms after layer L, vs REF
  layer        REF2         STD      SIMROW        INT8     ratio
      0   0.000e+00   1.041e-02   1.844e-02   2.106e-02       nan
      1   0.000e+00   1.632e-02   2.368e-02   2.740e-02       nan
      2   0.000e+00   9.332e-03   1.002e-02   1.091e-02       nan
      3   0.000e+00   1.479e-02   2.798e-02   3.113e-02       nan
      4   0.000e+00   1.538e-02   3.080e-02   3.345e-02       nan
      5   0.000e+00   1.590e-02   2.913e-02   3.199e-02       nan
      6   0.000e+00   1.268e-02   2.434e-02   2.728e-02       nan
      7   0.000e+00   1.652e-02   3.437e-02   3.922e-02       nan
      8   0.000e+00   1.569e-02   3.250e-02   3.742e-02       nan
      9   0.000e+00   1.645e-02   3.332e-02   3.765e-02       nan
     10   0.000e+00   1.411e-02   3.200e-02   3.631e-02       nan
     11   0.000e+00   1.383e-02   3.048e-02   3.919e-02       nan
     12   0.000e+00   1.475e-02   3.152e-02   3.921e-02       nan
     13   0.000e+00   1.531e-02   3.113e-02   3.779e-02       nan
     14   0.000e+00   1.594e-02   3.452e-02   4.045e-02       nan
     15   0.000e+00   1.659e-02   3.640e-02   4.364e-02       nan
     16   0.000e+00   1.660e-02   3.463e-02   4.219e-02       nan
     17   0.000e+00   1.661e-02   3.644e-02   4.354e-02       nan
     18   0.000e+00   1.375e-02   2.649e-02   3.009e-02       nan
     19   0.000e+00   1.786e-02   3.490e-02   3.674e-02       nan
     20   0.000e+00   1.603e-02   3.796e-02   3.723e-02       nan
     21   0.000e+00   2.011e-02   3.748e-02   4.286e-02       nan
     22   0.000e+00   1.967e-02   3.721e-02   4.171e-02       nan
     23   0.000e+00   2.076e-02   3.688e-02   4.144e-02       nan
     24   0.000e+00   2.000e-02   3.738e-02   4.276e-02       nan
     25   0.000e+00   2.012e-02   3.529e-02   3.959e-02       nan
     26   0.000e+00   1.781e-02   3.589e-02   4.020e-02       nan
     27   0.000e+00   1.888e-02   4.121e-02   4.452e-02       nan
     28   0.000e+00   2.063e-02   4.384e-02   4.798e-02       nan
     29   0.000e+00   1.902e-02   4.222e-02   4.545e-02       nan
     30   0.000e+00   1.883e-02   4.299e-02   4.734e-02       nan
     31   0.000e+00   1.978e-02   4.529e-02   5.050e-02       nan
     32   0.000e+00   2.235e-02   4.898e-02   5.034e-02       nan
     33   0.000e+00   2.033e-02   4.891e-02   5.124e-02       nan
     34   0.000e+00   1.979e-02   4.275e-02   4.485e-02       nan
     35   0.000e+00   1.964e-02   4.832e-02   4.817e-02       nan
     36   0.000e+00   2.034e-02   4.935e-02   4.807e-02       nan
     37   0.000e+00   2.126e-02   4.771e-02   4.963e-02       nan
     38   0.000e+00   2.042e-02   4.507e-02   4.694e-02       nan
     39   0.000e+00   1.966e-02   4.357e-02   4.437e-02       nan
     40   0.000e+00   1.913e-02   4.403e-02   4.521e-02       nan
     41   0.000e+00   2.149e-02   4.839e-02   4.770e-02       nan
     42   0.000e+00   2.090e-02   4.238e-02   4.291e-02       nan
     43   0.000e+00   1.828e-02   5.178e-02   4.872e-02       nan
     44   0.000e+00   2.006e-02   4.871e-02   4.910e-02       nan
     45   0.000e+00   1.842e-02   4.376e-02   4.500e-02       nan
     46   0.000e+00   1.829e-02   4.010e-02   4.353e-02       nan
     47   0.000e+00   1.991e-02   4.436e-02   4.499e-02       nan
     48   0.000e+00   2.315e-02   4.195e-02   4.696e-02       nan
     49   0.000e+00   1.812e-02   3.613e-02   3.824e-02       nan
     50   0.000e+00   1.540e-02   3.000e-02   3.423e-02       nan
     51   0.000e+00   1.618e-02   3.524e-02   3.798e-02       nan
     52   0.000e+00   1.636e-02   3.919e-02   3.463e-02       nan
     53   0.000e+00   1.545e-02   3.405e-02   3.572e-02       nan
     54   0.000e+00   1.501e-02   2.966e-02   3.281e-02       nan
     55   0.000e+00   1.713e-02   3.426e-02   3.962e-02       nan
     56   0.000e+00   1.890e-02   3.807e-02   3.914e-02       nan
     57   0.000e+00   1.994e-02   3.857e-02   4.373e-02       nan
     58   0.000e+00   1.770e-02   3.636e-02   5.194e-02       nan
     59   0.000e+00   1.522e-02   3.054e-02   3.237e-02       nan
     60   0.000e+00   1.796e-02   3.491e-02   3.855e-02       nan
     61   0.000e+00   1.729e-02   3.209e-02   3.774e-02       nan
     62   0.000e+00   1.563e-02   3.266e-02   3.277e-02       nan
     63   0.000e+00   2.090e-02   4.362e-02   6.301e-02       nan

  chunk 6: rel rms after layer L, vs REF
  layer        REF2         STD      SIMROW        INT8     ratio
      0   0.000e+00   1.309e-02   2.407e-02   2.297e-02       nan
      1   0.000e+00   1.127e-02   2.050e-02   2.610e-02       nan
      2   0.000e+00   7.867e-03   1.343e-02   2.427e-02       nan
      3   0.000e+00   1.249e-02   2.945e-02   2.894e-02       nan
      4   0.000e+00   1.495e-02   2.999e-02   3.616e-02       nan
      5   0.000e+00   1.744e-02   3.380e-02   3.908e-02       nan
      6   0.000e+00   1.482e-02   3.122e-02   3.794e-02       nan
      7   0.000e+00   1.674e-02   3.687e-02   4.222e-02       nan
      8   0.000e+00   1.600e-02   3.467e-02   4.138e-02       nan
      9   0.000e+00   1.495e-02   3.346e-02   3.740e-02       nan
     10   0.000e+00   1.340e-02   2.894e-02   3.334e-02       nan
     11   0.000e+00   1.617e-02   3.960e-02   4.308e-02       nan
     12   0.000e+00   1.472e-02   3.260e-02   3.774e-02       nan
     13   0.000e+00   1.455e-02   3.316e-02   3.982e-02       nan
     14   0.000e+00   1.374e-02   3.187e-02   3.517e-02       nan
     15   0.000e+00   1.446e-02   3.251e-02   3.794e-02       nan
     16   0.000e+00   1.451e-02   3.178e-02   3.616e-02       nan
     17   0.000e+00   1.554e-02   3.662e-02   3.852e-02       nan
     18   0.000e+00   1.477e-02   3.084e-02   3.584e-02       nan
     19   0.000e+00   1.470e-02   3.164e-02   3.603e-02       nan
     20   0.000e+00   1.481e-02   2.857e-02   3.148e-02       nan
     21   0.000e+00   1.616e-02   3.447e-02   3.978e-02       nan
     22   0.000e+00   1.424e-02   3.443e-02   3.765e-02       nan
     23   0.000e+00   1.507e-02   3.424e-02   3.812e-02       nan
     24   0.000e+00   1.571e-02   3.537e-02   4.027e-02       nan
     25   0.000e+00   1.527e-02   3.405e-02   4.018e-02       nan
     26   0.000e+00   1.536e-02   3.358e-02   3.928e-02       nan
     27   0.000e+00   1.740e-02   3.998e-02   4.611e-02       nan
     28   0.000e+00   2.001e-02   4.339e-02   4.816e-02       nan
     29   0.000e+00   1.829e-02   3.888e-02   4.424e-02       nan
     30   0.000e+00   2.050e-02   3.510e-02   4.370e-02       nan
     31   0.000e+00   2.386e-02   4.306e-02   5.176e-02       nan
     32   0.000e+00   2.215e-02   4.113e-02   5.035e-02       nan
     33   0.000e+00   2.643e-02   4.675e-02   6.222e-02       nan
     34   0.000e+00   3.682e-02   4.712e-02   8.082e-02       nan
     35   0.000e+00   6.778e-02   5.597e-02   1.540e-01       nan
     36   0.000e+00   6.521e-02   5.539e-02   1.547e-01       nan
     37   0.000e+00   6.758e-02   5.395e-02   1.578e-01       nan
     38   0.000e+00   7.343e-02   6.011e-02   1.698e-01       nan
     39   0.000e+00   9.330e-02   8.473e-02   2.089e-01       nan
     40   0.000e+00   8.745e-02   7.769e-02   2.025e-01       nan
     41   0.000e+00   8.709e-02   8.858e-02   2.088e-01       nan
     42   0.000e+00   7.826e-02   7.944e-02   1.796e-01       nan
     43   0.000e+00   9.709e-02   1.062e-01   1.954e-01       nan
     44   0.000e+00   9.266e-02   1.091e-01   1.887e-01       nan
     45   0.000e+00   8.477e-02   1.006e-01   1.865e-01       nan
     46   0.000e+00   9.827e-02   1.159e-01   2.283e-01       nan
     47   0.000e+00   9.936e-02   1.469e-01   2.805e-01       nan
     48   0.000e+00   9.428e-02   1.344e-01   4.140e-01       nan
     49   0.000e+00   1.048e-01   1.407e-01   5.638e-01       nan
     50   0.000e+00   1.191e-01   1.439e-01   7.070e-01       nan
     51   0.000e+00   1.017e-01   1.123e-01   7.308e-01       nan
     52   0.000e+00   9.491e-02   1.037e-01   6.947e-01       nan
     53   0.000e+00   9.770e-02   9.933e-02   7.521e-01       nan
     54   0.000e+00   1.161e-01   1.119e-01   8.309e-01       nan
     55   0.000e+00   1.081e-01   1.223e-01   7.901e-01       nan
     56   0.000e+00   9.248e-02   9.610e-02   6.860e-01       nan
     57   0.000e+00   6.786e-02   8.301e-02   8.499e-01       nan
     58   0.000e+00   1.275e-01   8.523e-02   9.230e-01       nan
     59   0.000e+00   8.650e-02   7.808e-02   6.447e-01       nan
     60   0.000e+00   1.180e-01   1.238e-01   8.824e-01       nan
     61   0.000e+00   1.266e-01   1.070e-01   9.617e-01       nan
     62   0.000e+00   1.004e-01   1.069e-01   8.205e-01       nan
     63   0.000e+00   7.856e-02   8.295e-02   6.734e-01       nan

PAIRED TV @ canonical (T=1.0, top_p=0.95, top_k=20), prefill positions.
Each row is A scored against B -- B is that ROW's reference, not the
table's.  TV is the expected speculative accept-rate loss, so the
column is read directly in percentage points.  The interval is a
cluster bootstrap over PROMPTS; the IQR is the raw spread over
positions with no floor applied to it.

  A       vs B        n  mean TV pp           95% CI pp    p25 pp    p50 pp    p75 pp    p95 pp    max pp
  REF     REF2      160      0.0000 [  0.0000,  0.0000]    0.0000    0.0000    0.0000    0.0000    0.0000
  REF     STD       160      1.5870 [  0.8665,  2.5456]    0.0000    0.2346    1.9388    4.4625   49.3853
  REF     SIMROW    160      2.0790 [  1.3515,  3.0505]    0.0000    0.4233    2.8217    7.2345   43.7824
  REF     INT8      160      2.6990 [  1.5135,  4.2872]    0.0000    0.6290    2.7828    6.7547   92.2893
  REF2    STD       160      1.5870 [  0.8665,  2.5456]    0.0000    0.2346    1.9388    4.4625   49.3853
  REF2    SIMROW    160      2.0790 [  1.3515,  3.0505]    0.0000    0.4233    2.8217    7.2345   43.7824
  REF2    INT8      160      2.6990 [  1.5135,  4.2872]    0.0000    0.6290    2.7828    6.7547   92.2893
  STD     SIMROW    160      2.7028 [  1.4994,  4.4587]    0.0000    0.5695    2.8896    7.6075   79.0139
  STD     INT8      160      2.9814 [  1.6831,  4.5573]    0.0000    0.5481    3.1110    8.6776   89.6264
  SIMROW  INT8      160      3.1098 [  1.6644,  4.9750]    0.0000    0.9935    3.3582    8.1179   95.3613

  TV distance matrix, mean TV pp (symmetric):
          REF      REF2       STD    SIMROW      INT8
  REF        0.0000    0.0000    1.5870    2.0790    2.6990
  REF2       0.0000    0.0000    1.5870    2.0790    2.6990
  STD        1.5870    1.5870    0.0000    2.7028    2.9814
  SIMROW     2.0790    2.0790    2.7028    0.0000    3.1098
  INT8       2.6990    2.6990    2.9814    3.1098    0.0000

  TV distance matrix, median TV pp (symmetric):
          REF      REF2       STD    SIMROW      INT8
  REF        0.0000    0.0000    0.2346    0.4233    0.6290
  REF2       0.0000    0.0000    0.2346    0.4233    0.6290
  STD        0.2346    0.2346    0.0000    0.5695    0.5481
  SIMROW     0.4233    0.4233    0.5695    0.0000    0.9935
  INT8       0.6290    0.6290    0.5481    0.9935    0.0000

lm_head CUT -- the divergence entering the head vs leaving it,
prefill positions, relative rms.  'in' is the residual stream after
24 layers; 'out' is the centred logits.  amp = out/in: at 1.0 the
head passes the body's divergence through unchanged, above it the
head adds its own.

  A       vs B        n   rel rms IN  rel rms OUT      amp
  REF     REF2      160   0.0000e+00   0.0000e+00      nan
  REF     STD       160   3.9650e-02   3.6420e-02    0.919
  REF     SIMROW    160   5.8675e-02   5.4017e-02    0.921
  REF     INT8      160   6.9933e-02   6.3770e-02    0.912
  REF2    STD       160   3.9650e-02   3.6420e-02    0.919
  REF2    SIMROW    160   5.8675e-02   5.4017e-02    0.921
  REF2    INT8      160   6.9933e-02   6.3770e-02    0.912
  STD     SIMROW    160   6.8578e-02   6.3376e-02    0.924
  STD     INT8      160   7.1067e-02   6.5211e-02    0.918
  SIMROW  INT8      160   8.2890e-02   7.5861e-02    0.915

PER-TOKEN AMAX -- is the crest factor of the quantised row what
selects the tail?  crest = amax/rms of the int8 GEMM's input row;
under the uniform-residual model the relative noise the row picks
up is crest/(127*sqrt(12)) = crest/440, and qerr is that noise
measured directly by round-tripping the scored row through int8.
arm supplying the statistics: INT8

  statistic                                  p50         p95         p99         max
  crest, scored row (max/linear)       1.174e+01   1.421e+01   1.667e+01   1.723e+01
  crest, any row in chunk              1.896e+01   2.112e+01   2.302e+01   2.330e+01
  qerr, scored row (rel rms)           2.661e-02   3.222e-02   3.759e-02   3.908e-02
  n positions: 160

What the per-token scale is paying for, over the whole census.
crest_g is the crest a per-128-group scale would leave --
sqrt(mean_b amax_b^2)/rms -- and its ratio to crest is the noise
reduction available.  conc is max/median of the per-128-block rms:
at 1 the row is flat and grouping buys nothing.
  statistic                                  p50         p95         p99         max
  crest_g (per-128 scale)                  2.602       2.878       2.935       3.160
  crest / crest_g (noise cut)              2.160       4.127       5.378       9.408
  block rms concentration                  3.315      12.039      24.118      77.319
  n (class, position) samples: 64000
  heaviest-block STABILITY per class (share of positions whose argmax block is that class's modal block):
    p05=0.037 p50=0.544 p95=1.000 over 400 classes
    a static outlier split needs this near 1.0; at 1/nblocks (~0.025 for K=5120) the heavy block is a different block every token and no static partition exists

The 10 worst INT8 positions, with the crest factor at each.  If the
per-token amax is the mechanism, these are the extreme-crest rows.
   prompt  chunk      INT8 KL   crest row  crest chunk        qerr  worst linear
       16      6    8.940e+00      11.362       19.089   2.552e-02  L21.linear_attn.out_proj
        0      7    2.040e+00      12.466       20.076   2.823e-02  L13.linear_attn.out_proj
        0      1    3.977e-01      11.198       18.043   2.551e-02  L8.linear_attn.out_proj
        2      7    5.191e-02      12.006       21.493   2.707e-02  L55.self_attn.o_proj
        2      3    2.703e-02      11.240       20.907   2.523e-02  L2.linear_attn.out_proj
        0      2    2.209e-02      16.673       20.897   3.759e-02  L50.linear_attn.out_proj
       18      7    2.161e-02       9.992       18.422   2.269e-02  L17.linear_attn.out_proj
       14      3    1.852e-02      10.895       18.788   2.467e-02  L13.linear_attn.out_proj
        3      4    1.831e-02      12.089       19.112   2.749e-02  L6.mlp.down_proj
       17      5    1.685e-02      11.056       20.079   2.519e-02  L51.self_attn.o_proj
  ... and the 10 MEDIAN INT8 positions, as the contrast:
       18      1    7.316e-04      11.799       20.011   2.661e-02  L0.mlp.down_proj
        2      4    7.954e-04      10.311       21.506   2.357e-02  L56.linear_attn.out_proj
       11      4    8.191e-04      11.021       20.534   2.501e-02  L0.mlp.down_proj
        5      7    8.734e-04      12.458       19.339   2.806e-02  L22.linear_attn.out_proj
       14      4    8.784e-04      10.265       19.919   2.328e-02  L6.linear_attn.out_proj
        1      4    9.407e-04      12.823       18.771   2.899e-02  L58.linear_attn.out_proj
       11      0    9.818e-04      10.919       18.473   2.477e-02  L2.mlp.down_proj
       11      7    9.858e-04      11.863       19.326   2.689e-02  L54.linear_attn.out_proj
       16      2    9.914e-04      14.086       18.511   3.200e-02  L51.mlp.down_proj
       16      1    1.031e-03      11.255       17.735   2.561e-02  L0.mlp.down_proj

Is the tail MORE NOISE or MORE RESPONSE?  rel rms of the residual
stream entering lm_head, per position, and the INT8/STD ratio.
Flat ratio across the tail => the position amplifies every arm and
int8 is merely 2.4x noisier everywhere.  Ratio spiking on the tail
=> int8 injects locally more noise there, which is a mechanism.
  ratio over ALL 160 positions: p50=2.013 p95=3.197 max=8.513
  ratio at the 10 worst INT8 positions: min=0.239 median=2.147 max=7.908
   prompt  chunk      INT8 KL  relIN INT8   relIN STD    ratio
       16      6    8.940e+00  8.3688e-01  1.0582e-01    7.908
        0      7    2.040e+00  7.0663e-01  4.5197e-01    1.563
        0      1    3.977e-01  3.2760e-01  5.1640e-01    0.634
        2      7    5.191e-02  1.0871e+00  3.0422e-01    3.574
        2      3    2.703e-02  1.5538e-01  6.4976e-01    0.239
        0      2    2.209e-02  1.5306e-01  1.1289e-01    1.356
       18      7    2.161e-02  4.8145e-02  2.0128e-02    2.392
       14      3    1.852e-02  6.7665e-02  2.1165e-02    3.197
        3      4    1.831e-02  4.9491e-02  1.8847e-02    2.626
       17      5    1.685e-02  7.8604e-02  4.1313e-02    1.903
  Spearman(relIN ratio, INT8 KL) = 0.177   Spearman(relIN INT8, INT8 KL) = 0.105

Rank correlation of the crest factor with each arm's KL.  STD is the
NULL CONTROL: it runs no quantiser, so the crest factor cannot cause
its KL.  A crest that ranks STD as well as it ranks INT8 is ranking
hard POSITIONS, not the quantiser, and explains nothing.
  crest statistic                    REF       REF2        STD     SIMROW       INT8  INT8 excess
  crest, scored row               -0.035     -0.035     -0.014      0.070     -0.006        0.073
  crest, any row                  -0.200     -0.200      0.147      0.215      0.185        0.157
  qerr, scored row                -0.032     -0.032     -0.019      0.067     -0.008        0.076

Mean crest of the scored row, by linear class -- the 15 heaviest.
A class well above the Gaussian-rotation expectation (~4.5 for a
128-wide Hadamard over K=5120) is a class the per-token scale is
wasting levels on.
  linear class                                        mean crest         p95         max
  L2.mlp.down_proj                                         8.768      11.681      13.547
  L2.linear_attn.out_proj                                  8.288      11.549      13.587
  L0.mlp.down_proj                                         8.151      11.632      12.463
  L58.linear_attn.out_proj                                 8.064      13.624      17.229
  L4.linear_attn.out_proj                                  7.805      11.044      13.454
  L5.linear_attn.out_proj                                  7.688      11.693      14.842
  L1.linear_attn.out_proj                                  7.294      10.832      13.080
  L17.linear_attn.out_proj                                 7.107      10.029      13.344
  L56.linear_attn.out_proj                                 7.080      11.498      15.215
  L49.linear_attn.in_proj_z                                7.065       7.958       8.256
  L44.linear_attn.in_proj_z                                7.028       7.998       8.611
  L45.linear_attn.in_proj_z                                7.000       7.781       8.154
  L6.linear_attn.out_proj                                  6.999      10.021      11.511
  L61.linear_attn.out_proj                                 6.989      10.141      12.757
  L46.linear_attn.in_proj_z                                6.981       7.745       8.087
