Verbosity ≠ Veracity: Demystify Verbosity Compensation Behavior of Large Language Models
https://arxiv.org/abs/2411.07858
Verbosity Compensation" (VC)
回答に自信がないとき、人間が一部でも当たっていればと必要以上に多く述べてしまう事象はLLMにも起きる
We propose a simple yet effective cascade algorithm that replaces the verbose responses with the other model-generated responses. (Abstract)
https://github.com/psunlpgroup/VerbosityLLM
Figure 1にVCの2事例
質問の繰り返し
曖昧な回答