{% extends "base.html" %} {% block title %}Benchmark Results - lm-eval-ledger{% endblock %} {% block content %}

Leaderboard

Tasks best per model
{% for t in tasks %} {{ t.task }} best {{ "%.3f"|format(t.best or 0) }} · {{ t.n }} entr{{ "ies" if t.n > 1 else "y" }} {% endfor %}

Benchmark results

{% for b in rows %} {% endfor %}
runmodeltask (few-shot) acc nno answertimetok/s
{{ b.run_id }} {{ b.model_tag }} {{ b.task }}{% if b.fewshot_k is not none %}({{ b.fewshot_k }}){% endif %} {{ medals.get(b.benchmark_id, "") }}{% if b.verified_accuracy is not none %} {{ "%.3f"|format(b.verified_accuracy) }} {% else %}{{ "%.3f"|format(b.accuracy) if b.accuracy is not none else "…" }}{% endif %} {{ b.total_examples }} {{ b.no_answer_count }} {{ b.duration_seconds|fmtdur if b.duration_seconds else "-" }} {{ "%.0f"|format(b.gen_tokens / b.gen_seconds) if b.gen_tokens and b.gen_seconds else "-" }}
{% endblock %}