Tag: sandbagging
All the articles with the tag "sandbagging".
- 4.0
Below-Chance Blindness: Prompted Underperformance in Small LLMs Produces Positional Bias Rather than Answer Avoidance
用临床伪装检测逻辑检测 LLM 故意 underperformance(sandbagging)
All the articles with the tag "sandbagging".
用临床伪装检测逻辑检测 LLM 故意 underperformance(sandbagging)