Focus Directions Make Your Language Models Pay More Attention to Relevant Contexts
This addresses a key limitation in long-context LLMs for users needing reliable performance in tasks with extensive inputs, though it is incremental as it builds on existing attention mechanisms.
The paper tackles the problem of long-context large language models (LLMs) being distracted by irrelevant contexts, and finds that increasing attention to relevant contexts via focus directions mitigates this issue, improving task alignment.
Long-context large language models (LLMs) are prone to be distracted by irrelevant contexts. The reason for distraction remains poorly understood. In this paper, we first identify the contextual heads, a special group of attention heads that control the overall attention of the LLM. Then, we demonstrate that distraction arises when contextual heads fail to allocate sufficient attention to relevant contexts and can be mitigated by increasing attention to these contexts. We further identify focus directions, located at the key and query activations of these heads, which enable them to allocate more attention to relevant contexts without explicitly specifying which context is relevant. We comprehensively evaluate the effect of focus direction on various long-context tasks and find out focus directions could help to mitigate the poor task alignment of the long-context LLMs. We believe our findings could promote further research on long-context LLM alignment.