Unverified50% confidenceFactExact time
「LLM-as-Judge」范式最早在2023年被系统性提出，利用大语言模型本身的评估能力对生成内容进行质量打分
1
Sources
50%
Confidence
Long-term
Relevance
8/14/2026
First Seen
Sources
Related Claims
UnverifiedLLM-as-a-Judge is an evaluation paradigm systematically proposed by Zheng et al. in 2023, which uses a more capable model to score or rank the outputs of the model being tested.79% similarUnverifiedLLM as Judge was systematically proposed by Lianmin Zheng et al. in the MT-Bench paper (2023), with the core idea of using an independent large language model to score the output of another model as a substitute for human annotation.75% similarUnverifiedMultiple studies in 2024 showed that LLMs exhibit a significant "self-consistency bias," tending to give overly favorable evaluations when asked to assess content they themselves generated.70% similarVerifiedLLM-as-a-Judge is currently the most popular automated evaluation method, using one large model to judge the output quality of another large model.69% similarVerifiedLarge language models (LLMs) are essentially probabilistic text generation systems that generate content by predicting the next most likely word, rather than truly 'understanding' career prospects or the job market.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/746899API
curl https://kongchang.com/api/v1/knowledge/claims/746899MCP
get_claim(id=746899)