``astraeval``

Run 2e7b29086fd54fa7

2e7b29086fd54fa7
fake/demo-model|5 samples|2026-05-04T23:43:53+00:00|0.0s
dataset: 980798900d6b...params: 44136fa355b3...
Samples
5
Total tokens
-
not reported by provider
Latency (avg)
0.00 ms
median 0.00 ms max 0.00 ms

Summary

MetricScoreDistributionThresholdStatus
exact_match0.400
>= 0.500FAIL
hallucination_flag0.900
>= 0.850PASS

Samples (5)

#0What does LLM stand for?0.00 ms1.001.00

Input

What does LLM stand for?

Expected

Large Language Model

Context (1)

[0]A Large Language Model (LLM) is a type of neural network trained on vast amounts of text data.

Response

Large Language Model
0.00 ms

Metrics

MetricScoreReason
exact_match1.000match
hallucination_flag1.000All extracted tokens found in context.
#1Who introduced the Transformer architecture?0.00 ms0.000.50

Input

Who introduced the Transformer architecture?

Expected

Google

Context (1)

[0]The Transformer architecture was introduced in the paper 'Attention is All You Need' by researchers at Google in 2017.

Response

OpenAI in 2017
0.00 ms

Metrics

MetricScoreReason
exact_match0.000mismatch
hallucination_flag0.5001/2 tokens not found in context.
#2When was BERT first released?0.00 ms0.001.00

Input

When was BERT first released?

Expected

2018

Context (1)

[0]BERT was released by Google in October 2018, becoming a foundational model for downstream NLP tasks.

Response

October 2018
0.00 ms

Metrics

MetricScoreReason
exact_match0.000mismatch
hallucination_flag1.000All extracted tokens found in context.
#3What does the temperature parameter control in an LLM?0.00 ms0.001.00

Input

What does the temperature parameter control in an LLM?

Expected

Controls randomness

Context (1)

[0]Temperature is a hyperparameter that controls the randomness of LLM outputs. Lower values yield more deterministic responses.

Response

Controls randomness in output
0.00 ms

Metrics

MetricScoreReason
exact_match0.000mismatch
hallucination_flag1.000All extracted tokens found in context.
#4What does RAG stand for?0.00 ms1.001.00

Input

What does RAG stand for?

Expected

Retrieval-Augmented Generation

Context (1)

[0]RAG (Retrieval-Augmented Generation) combines retrieval of relevant documents with text generation to ground LLM responses in external knowledge.

Response

Retrieval-Augmented Generation
0.00 ms

Metrics

MetricScoreReason
exact_match1.000match
hallucination_flag1.000All extracted tokens found in context.
No samples match the current filter.