Across 51 quantized checkpoints, quality metrics fail to predict safety drops in 36 pairings and 10 hidden-danger cases, while a new RTSI screen routes all 10 dangerous rows to testing at matched bucket size.
How post-training reshapes LLMs: A mechanistic view on knowledge, truthfulness, refusal, and confidence.arXiv preprint arXiv:2504.02904, 2025
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 3roles
background 1polarities
unclear 1representative citing papers
Reasoning in large output spaces proceeds via shortlisting then fine-grained reasoning; this characterization enables a mechanistic distillation strategy that outperforms standard distillation.
citing papers explorer
-
Quality Is Not a Safety Proxy Under Quantization
Across 51 quantized checkpoints, quality metrics fail to predict safety drops in 36 pairings and 10 hidden-danger cases, while a new RTSI screen routes all 10 dangerous rows to testing at matched bucket size.
-
Characterize Then Distill: Mechanistic Reasoning in Large Output Spaces
Reasoning in large output spaces proceeds via shortlisting then fine-grained reasoning; this characterization enables a mechanistic distillation strategy that outperforms standard distillation.
- Same Targets, Different Computation: How Post-Training Divides Work Across Model Layers