When Expert Disagreement Hurts: Prestige and Selectivity in LLM Decision Revision

Research Hugging Face Papers Source language: Chinese

论文摘要(作者报告,未经本站复测):When language models revise a decision after another agent disagrees, neither reversal counts nor final accuracy shows whether they are fixing mistakes or abandoning correct answers.

Excerpt

论文摘要(作者报告,未经本站复测):When language models revise a decision after another agent disagrees, neither reversal counts nor final accuracy shows whether they are fixing mistakes or abandoning correct answers.

Short excerpt from the source — open the original for the full text.

Sources

← Back to the daily brief