repoGitHubTrust 82 · PrimaryPublished 5d agoLive · 5d ago

DaoyuanLi2816/pairjudge

Pairwise LLM judges (A/B/tie): budget-aware multi-turn packing, position-bias correction, pseudo-label distillation. Generalized from the 4th-place (gold) solution to Kaggle LMSYS Chatbot Arena.

Lineage graph

Paper → model → repo connections mined from source citations (Tier-1 exact match).

Why these links exist

Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.

PossiblePossibly related (embedding) · 49%When the Judge Changes, So Does the Measurement: Auditing LLM-as-Judge Reliability →
PossiblePossibly related (embedding) · 47%LLMs are stuck in a groupthink rut. This startup is trying to get them out. →
PossiblePossibly related (embedding) · 47%LLMs are stuck in a groupthink groove. This startup is trying to get them out. →
PossiblePossibly related (embedding) · 46%Evaluating long-term memory limits in stateless LLM chatbots — feedback needed [D] →
PossiblePossibly related (embedding) · 45%Large Tabular Models Excel Where LLMs Fail →

Implements

paperWhen the Judge Changes, So Does the Measurement: Auditing LLM-as-Judge Reliability

Covers

newsLLMs are stuck in a groupthink rut. This startup is trying to get them out.newsLLMs are stuck in a groupthink groove. This startup is trying to get them out.newsEvaluating long-term memory limits in stateless LLM chatbots — feedback needed [D]newsLarge Tabular Models Excel Where LLMs Fail

Related across the graph

newsLarge Tabular Models Excel Where LLMs Fail newsLLMs are stuck in a groupthink rut. This startup is trying to get them out.newsEvaluating long-term memory limits in stateless LLM chatbots — feedback needed [D]newsLLMs are stuck in a groupthink groove. This startup is trying to get them out.paperWhen the Judge Changes, So Does the Measurement: Auditing LLM-as-Judge Reliability

Topics