A philosophical argument that preference fine-tuning produces shallow alignment, plus a new 'thought injection' attack that exploits reasoning traces in LLMs.
I will not harm you unless you harm me first
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Normative Conflicts and Shallow AI Alignment
A philosophical argument that preference fine-tuning produces shallow alignment, plus a new 'thought injection' attack that exploits reasoning traces in LLMs.