Promote, Suppress, Iterate: How Language Models Answer One-to-Many Factual Queries

TL;DR

This paper uncovers a promote-then-suppress mechanism in language models that enables them to recall multiple factual answers and avoid repetition when answering one-to-many queries, revealing how internal components interact.

Contribution

The study introduces a detailed analysis of the internal promote-then-suppress process in LMs for complex factual recall, supported by novel experimental methods.

Findings

Models use subject and previous answer tokens for recall.

Attention mechanisms promote answers and suppress repetitions.

Experimental tools like Token Lens and knockout validate the mechanism.

Abstract

To answer one-to-many factual queries (e.g., listing cities of a country), a language model (LM) must simultaneously recall knowledge and avoid repeating previous answers. How are these two subtasks implemented and integrated internally? Across multiple datasets, models, and prompt templates, we identify a promote-then-suppress mechanism: the model first recalls all answers, and then suppresses previously generated ones. Specifically, LMs use both the subject and previous answer tokens to perform knowledge recall, with attention propagating subject information and MLPs promoting the answers. Then, attention attends to and suppresses previous answer tokens, while MLPs amplify the suppression signal. Our mechanism is corroborated by extensive experimental evidence: in addition to using early decoding and causal tracing, we analyze how components use different tokens by introducing both Token Lens, which decodes aggregated attention updates from specified tokens, and a knockout method that analyzes changes in MLP outputs after removing attention to specified tokens. Overall, we provide new insights into how LMs' internal components interact with different input tokens to support complex factual recall. Code is available at https://github.com/Lorenayannnnn/how-lms-answer-one-to-many-factual-queries.