https://arxiv.org/abs/2304.11111
Inducing anxiety in large language models increases exploration and bias (Julian Coda-Forno, Kristin Witte, Akshay K. Jagadish, Marcel Binz, Zeynep Akata, Eric Schulz)
아니...이게 대체 뭐죠? GPT-3.5에게 불안 지수 평가 질문지로 불안 지수를 평가해봤더니 사람보다 불안 지수가 평균적으로 더 높더라, 프롬프트로 경험하는 정서 상태를 바꿀 수 있는지 평가해봤더니 어느 정도 그렇더라, 그리고 이 불안 지수가 실제 모델의 패턴에 영향을 미치는지 알기 위해 2 armed bandit으로 테스트해봤더니 exploration-exploitation에 대한 행동이 바뀌더라...를 테스트해봤네요.
#llm