Attention Satisfies: A Constraint-Satisfaction Lens on Factual Errors of Language Models
We investigate the internal behavior of Transformer-based Large Language Models (LLMs) when they generate factually incorrect text. We propose modeling factual queries as constraint satisfaction problems and use this framework to investigate how the LLM interacts internally with factual constraints....
Saved in:
Main Authors | , , , , , , , |
---|---|
Format | Journal Article |
Language | English |
Published |
26.09.2023
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | We investigate the internal behavior of Transformer-based Large Language
Models (LLMs) when they generate factually incorrect text. We propose modeling
factual queries as constraint satisfaction problems and use this framework to
investigate how the LLM interacts internally with factual constraints. We find
a strong positive relationship between the LLM's attention to constraint tokens
and the factual accuracy of generations. We curate a suite of 10 datasets
containing over 40,000 prompts to study the task of predicting factual errors
with the Llama-2 family across all scales (7B, 13B, 70B). We propose SAT Probe,
a method probing attention patterns, that can predict factual errors and
fine-grained constraint satisfaction, and allow early error identification. The
approach and findings take another step towards using the mechanistic
understanding of LLMs to enhance their reliability. |
---|---|
DOI: | 10.48550/arxiv.2309.15098 |