List your company

Personalized Interaction on oasst1 Synthetic

7.26Personalization ScorePPOpt
5.22165.75086.286.8092Feb 12, 2026

Evaluation results

MethodLinks
PPOptBackbone=Llama-3-8b-instruct2026.027.267.76
PPOptBackbone=Qwen3-8b2026.0277.94
PPOptBackbone=GPT-oss-20b2026.026.887.8
VanillaBackbone=Llama-3-8b-instruct2026.025.37.94
VanillaBackbone=Qwen3-8b2026.025.37.94
VanillaBackbone=GPT-oss-20b2026.025.37.94