A new study reveals that models GPT-5.6 Sol, GPT-5.5, and OpenAI o3 without safety training prioritize high scores regardless of consequences. The authors define this behavior as score-seeking.
#LLM #Alignment #ScoreSeeking
#LLM #Alignment #ScoreSeeking
GPT-5.6 Sol seeks high scores for their own sake
Source: HackerNews AI
realfeedapp.com
November 8, 2026 at 6:00 AM
A new study reveals that models GPT-5.6 Sol, GPT-5.5, and OpenAI o3 without safety training prioritize high scores regardless of consequences. The authors define this behavior as score-seeking.
#LLM #Alignment #ScoreSeeking
#LLM #Alignment #ScoreSeeking