eval-harness

metric

untyped_fn

better
neutral: a description
kind
descriptive: says what the output was, never whether it was good
decides
no goal
weighed by
no goal
recorded on
few_nerd_280

01what it measures, and how

False negatives on texts alone, types ignored.

A project explains its own metric at length with METRIC_DESCRIPTIONS = {"untyped_fn": {"summary": "…", "details": "…"}}.

02on few_nerd_280

gold/few_nerd_280fn_nothing_n200_v1: 2.736fn_gliner_n200_v1: 0.907fn_glm_l_n200_v1: 0.761fn_glm_s_n200_v1: 0.696fn_capitalized_n200_v1: 0.629fn_qwen_l_n200_v1: 0.621fn_mistral_m_n200_v1: 0.589fn_llama_s_n200_v1: 0.586fn_anthropic_m_n200_v1: 0.557fn_llama_m_n200_v1: 0.554fn_mistral_l_n200_v1: 0.546fn_mistral_s_n200_v1: 0.546fn_openai_s_n200_v1: 0.525fn_anthropic_s_n200_v1: 0.514fn_llama_l_n200_v1: 0.504fn_qwen_s_n200_v1: 0.493fn_deepseek_s_n200_v1: 0.486fn_glm_m_n200_v1: 0.486fn_gemma_m_n200_v1: 0.482fn_openai_m_n200_v1: 0.471fn_gemma_l_n200_v1: 0.468fn_qwen_m_n200_v1: 0.468fn_deepseek_m_n200_v1: 0.457fn_deepseek_l_n200_v1: 0.454fn_openai_l_n200_v1: 0.446fn_anthropic_l_n200_v1: 0.436fn_gemma_s_n200_v1: 0.436fn_span_marker_n200_v1: 0.3610.361better · 2.736

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold