Llama See, Llama Do: A Mechanistic Perspective on Contextual Entrainment and Distraction in LLMs

ACL ARR 2025 February Submission1154 Authors

Abstract:

We observe a novel phenomenon, *contextual entrainment*, across a wide range of language models (LMs) and prompt settings, providing a new mechanistic perspective on how LMs become distracted by “irrelevant” contextual information in the input prompt. Specifically, LMs assign significantly higher logits (or probabilities) to any tokens that have previously appeared in the context prompt, even for random tokens. This suggests that contextual entrainment is a *mechanistic* phenomenon, occurring independently of the relevance or semantic relation of the tokens to the question or the rest of the sentence. We find statistically significant evidence that the magnitude of contextual entrainment is influenced by semantic factors. Counterfactual prompts have a greater effect compared to factual ones, suggesting that while contextual entrainment is a mechanistic phenomenon, it is modulated by semantic factors.

We hypothesize that a cluster of attention heads — the *entrainment heads* — corresponds to contextual entrainment. Using a novel entrainment head discovery method based on differentiable masking, we identify these heads across various settings. When we “turn off” these heads, i.e., set their output to zero, the effect of contextual entrainment is significantly attenuated, causing the model to generate output that capitulates to what it would produce if no distracting context were provided. Our discovery of contextual entrainment, along with our investigation into LM distraction via the entrainment heads, marks a key step towards mechanistic analysis and mitigation of the distraction problem.

Paper Type: Long

Research Area: Interpretability and Analysis of Models for NLP

Research Area Keywords: Distraction, Contextual Entrainment, Attention Heads, RAG, Differentiable Masking

Contribution Types: Model analysis & interpretability, Publicly available software and/or pre-trained models

Languages Studied: English

Submission Number: 1154