Search papers, labs, and topics across Lattice.
Affiliation:
3
0
7
8
Penalizing shifts in safety representations during reasoning fine-tuning can restore LLM safety without sacrificing performance, revealing a critical interplay between reasoning and safety in model training.
MemCatalyst reveals that targeted data poisoning can drastically boost membership inference accuracy in Vision-Language Models with minimal resource expenditure.
LLMs possess a "word recovery" mechanism that allows them to reconstruct canonical word-level tokens from character-level inputs, explaining their surprising robustness to non-canonical tokenization.
