Using a Markov-chain model of LLM fine-tuning, the paper claims a training-step bound that makes refusal states absorbing and proposes ensemble width as a substitute for alignment depth, but the proof is flawed.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Safety Alignment Depth in Large Language Models: A Markov Chain Perspective
Using a Markov-chain model of LLM fine-tuning, the paper claims a training-step bound that makes refusal states absorbing and proposes ensemble width as a substitute for alignment depth, but the proof is flawed.