Sentiment is represented as a single linear direction in LLM activation space that is causally relevant across tasks and is summarized at punctuation and names in addition to charged words.
Inducing Causal Structure for Interpretable Neural Networks , url =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2023 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Linear Representations of Sentiment in Large Language Models
Sentiment is represented as a single linear direction in LLM activation space that is causally relevant across tasks and is summarized at punctuation and names in addition to charged words.