ResearchBenchmarksPersonalized Interaction on AI2 ARC SyntheticFollow7.38Personalization ScorePPOpt4.50965.254866.7452Feb 12, 2026Evaluation ResultsMethodMethodLinksPersonalization ScoreTask Completion ScorePPOptBackbone=Llama-3-8b-in...Backbone=Llama-3-8b-instruct2026.027.389.5PPOptBackbone=GPT-oss-20bBackbone=GPT-oss-20b2026.027.349.48PPOptBackbone=Qwen3-8bBackbone=Qwen3-8b2026.026.69.3VanillaBackbone=Llama-3-8b-in...Backbone=Llama-3-8b-instruct2026.024.629.56VanillaBackbone=Qwen3-8bBackbone=Qwen3-8b2026.024.629.56VanillaBackbone=GPT-oss-20bBackbone=GPT-oss-20b2026.024.629.56