eval-harness

metric

precision

better
higher
kind
quality: measured against a reference or by a judge
decides
no goal
weighed by
no goal
recorded on
few_nerd_280

01what it measures, and how

Typed precision: the share of the predicted entities that match a gold one in text and type.

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

02on few_nerd_280

gold/few_nerd_280fn_span_marker_n200_v1: 0.773fn_openai_m_n200_v1: 0.675fn_anthropic_l_n200_v1: 0.674fn_anthropic_m_n200_v1: 0.672fn_gemma_m_n200_v1: 0.67fn_openai_l_n200_v1: 0.66fn_qwen_s_n200_v1: 0.653fn_gemma_l_n200_v1: 0.641fn_qwen_m_n200_v1: 0.634fn_qwen_l_n200_v1: 0.611fn_mistral_l_n200_v1: 0.587fn_deepseek_m_n200_v1: 0.58fn_llama_l_n200_v1: 0.579fn_openai_s_n200_v1: 0.571fn_deepseek_l_n200_v1: 0.567fn_glm_m_n200_v1: 0.565fn_anthropic_s_n200_v1: 0.564fn_mistral_m_n200_v1: 0.561fn_llama_m_n200_v1: 0.553fn_llama_s_n200_v1: 0.551fn_deepseek_s_n200_v1: 0.548fn_glm_l_n200_v1: 0.546fn_glm_s_n200_v1: 0.539fn_gemma_s_n200_v1: 0.511fn_mistral_s_n200_v1: 0.488fn_gliner_n200_v1: 0.436fn_capitalized_n200_v1: 0.176fn_nothing_n200_v1: 0.1250.125better · 0.773

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