Refusal-Gated Decoding: Preserving Refusal Behavior Under High-Temperature Sampling
Quick summary
arXiv:2607.20791v2 Announce Type: replace Abstract: Recent advances in truncation-based sampling have helped mitigate drawbacks of high-temperature sampling such as neural text degeneration, thereby enabling greater diversity without sacrificing coherence. However, increasing the entropy of the token probability distribution via high temperatures has also been shown to weaken the model's refusal response. Existing solutions for maintaining the refusal behavior of LLMs either replace the model's own refusal decision with a separate safety classifier or alter its output distribution for every pr
Key takeaways
- arXiv:2607.20791v2 Announce Type: replace Abstract: Recent advances in truncation-based sampling have helped mitigate drawbacks of high-temperature sampling such as neural text degeneration, thereby enabling greater diversity without sacrificing coherence.
- However, increasing the entropy of the token probability distribution via high temperatures has also been shown to weaken the model's refusal response.
- Existing solutions for maintaining the refusal behavior of LLMs either replace the model's own refusal decision with a separate safety classifier or alter its output distribution for every pr
Why it matters
“Refusal-Gated Decoding: Preserving Refusal Behavior Under High-Temperature Sampling” shows why AI risk cannot be reduced to answer accuracy. Access controls, logging, human approval and incident response need to be designed into the workflow from the start.
Member comments