List your company

Session-level Correlation on MT-Bench

0.74Spearman Correlation (ρ)SKG-Eval
0.480.54750.6150.6825May 15, 2026

Evaluation results

MethodLinks
SKG-Eval2026.050.74—0.56
SKG-Eval2026.050.730.76—
GPT-4o-Judgemode=history-aware2026.050.69—0.51
GPT-4o-JudgeType=history-aware2026.050.660.69—
GPT-4o-Judgemode=turn-only2026.050.63—0.47
DeepEvalmodel=GPT-4o2026.050.61—0.45
DeepEval + GPT-4oEvaluator=DeepEval + GPT-4o2026.050.590.62—
GPT-4o-JudgeType=turn-only2026.050.570.6—
LLM-Eval2026.050.54—0.39
LLM-Eval2026.050.530.56—
ECoh2026.050.51—0.37
ECoh2026.050.490.52—