Trust or Escalate: LLM Judges with Provable Guarantees for Human Agreement
원문 초록
We present a principled approach to provide LLM-based evaluation with a rigorous guarantee of human agreement. We first propose that a reliable evaluation method should not uncritically rely on model preferences for pairwise evaluation, but rather assess the confidence of judge models and selectively decide when to trust its judgement. We then show that under this selective evaluation framework, human agreement can be provably guaranteed -- such that the model evaluation aligns with that of humans to a user-specified agreement level. As part of our framework, we also introduce Simulated Annotators, a novel confidence estimation method that significantly improves judge calibration and thus enables high coverage of evaluated instances. Finally, we propose Cascaded Selective Evaluation, where we use cheaper models as initial judges and escalate to stronger models only when necessary -- again, while still providing a provable guarantee of human agreement. Experimental results show that Cascaded Selective Evaluation guarantees strong alignment with humans, far beyond what LLM judges could achieve without selective evaluation. For example, on a subset of Chatbot Arena where GPT-4 almost never achieves 80% human agreement, our method, even while employing substantially cost-effective models such as Mistral-7B, guarantees over 80% human agreement with almost 80% test coverage.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Generalized Adaptive Transfer Network: Enhancing Transfer Learning in Reinforcement Learning Across Domains
MEDAL: A Framework for Benchmarking LLMs as Multilingual Open-Domain Chatbots and Dialogue Evaluators
Is Your LLM Overcharging You? Tokenization, Transparency, and Incentives
A Fully Generative Motivational Interviewing Counsellor Chatbot for Moving Smokers Towards the Decision to Quit
SWE-Dev: Evaluating and Training Autonomous Feature-Driven Software Development
X-MAS: Towards Building Multi-Agent Systems with Heterogeneous LLMs
토론
이 논문과 연결된 포럼 글이 아직 없습니다.