List your company

Negotiation on Student (held-out test)

100Success RateVanilla (LLM)
9597.5100102.5May 26, 2026

Evaluation results

MethodLinks
Vanilla (LLM)Backbone=Qwen3.5-Plus2026.0510045.92.5
Vanilla (SLM)Backbone=Qwen2.5-7B-Instruct2026.05100153.4
Random2026.0510043.82.5
IQLBackbone=Qwen3.5-Plus2026.0510047.92.8
IQL+SFTBackbone=Qwen2.5-7B-Instruct2026.0510051.72.5
IQL+JPOBackbone=Qwen2.5-7B-Instruct2026.0510023.55.5
IQL+SFT+JPOBackbone=Qwen2.5-7B-Instruct, Note=Full EmoDistill pipeline2026.0510052.63.1