Safety, Alignment & InterpretabilityUCL, NYU, Anthropic (Khan et al.)New York University (NYU)Anthropic
Debating with More Persuasive LLMs (ICML 2024 Best Paper)
AIMade by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made
Debating with More Persuasive LLMs (ICML 2024 Best Paper)
Why it matters
The first empirical evidence that debate-based oversight actually raises the accuracy of non-expert judges
Source — Primary sources first; secondary reports are labelled “Press”
Paperarxiv.orghttps://arxiv.org/abs/2402.06782Added in Week 35, 2026 · See that week
On the map — ⑩ Safety, Alignment line
- Previous stopWeak-to-strong generalization
- This recordDebating with More Persuasive LLMs (ICML 2024 Best Paper)
- Next stopfirst Model Spec