RM-Bench ↗68.3%
internlm2-20b-reward · RM-Bench · Overall accuracy · official leaderboard ↗
Reported model / MODEL EVIDENCE
internlm2-20b-reward
3 benchmarks · 3 observations · 3 sources
Scores at a glance
3 benchmarks · large dot: internlm2-20b-reward; small dots: closest models where reported
JudgeBench ↗63.43%
InternLM2-20B-Reward · JudgeBench · Overall pairwise accuracy · Reward models ↗
internLM2-20B-Reward · PPE (Preference Proxy Evaluations) · Correctness accuracy · mean · correctness suite ↗
Closest models by score
Similar positions on shared benchmark leaderboards. Smaller gaps are closer.
3 of 3 shared benchmark comparisons
3 of 3 shared benchmark comparisons
How similarity is measured
Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.
Sources
Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.
- JudgeBench: A Benchmark for Evaluating LLM-based Judges — published result table ↗
All 8 authors
Sijun Tan, Siyuan Zhuang, Kyle Montgomery, William Y. Tang, Alejandro Cuadron, Chenguang Wang, Raluca Ada Popa, Ion Stoica
- How to Evaluate Reward Models for RLHF — published result table ↗
All 9 authors
Evan Frick, Tianle Li, Connor Chen, Wei-Lin Chiang, Anastasios N. Angelopoulos, Jiantao Jiao, Banghua Zhu, Joseph E. Gonzalez, Ion Stoica
- RM-Bench: Benchmarking Reward Models of Language Models with Subtlety and Style — published result table ↗
Names used in source tables
InternLM2-20B-Reward · internLM2-20B-Reward · internlm2-20b-reward