By re-sampling alternate tokens at each position of generated text, the authors find many 'forking tokens' where a single token choice flips the model's final answer distribution, including chain-of-thought steps and punctuation-like tokens.
Terry loves to swim but he never learned how to dive
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Forking Paths in Neural Text Generation
By re-sampling alternate tokens at each position of generated text, the authors find many 'forking tokens' where a single token choice flips the model's final answer distribution, including chain-of-thought steps and punctuation-like tokens.