Spotlight Your Instructions: Instruction-following with Dynamic Attention Steering
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866908784059744256 |
|---|---|
| author | Venkateswaran, Praveen Contractor, Danish |
| author_facet | Venkateswaran, Praveen Contractor, Danish |
| contents | In many real-world applications, users rely on natural language instructions to guide large language models (LLMs) across a wide range of tasks. These instructions are often complex, diverse, and subject to frequent change. However, LLMs do not always attend to these instructions reliably, and users lack simple mechanisms to emphasize their importance beyond modifying prompt wording or structure. To address this, we present an inference-time method that enables users to emphasize specific parts of their prompt by steering the model's attention toward them, aligning the model's perceived importance of different prompt tokens with user intent. Unlike prior approaches that are limited to static instructions, require significant offline profiling, or rely on fixed biases, we dynamically update the proportion of model attention given to the user-specified parts--ensuring improved instruction following without performance degradation. We demonstrate that our approach improves instruction following across a variety of tasks involving multiple instructions and generalizes across models of varying scales. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_12025 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Spotlight Your Instructions: Instruction-following with Dynamic Attention Steering Venkateswaran, Praveen Contractor, Danish Machine Learning In many real-world applications, users rely on natural language instructions to guide large language models (LLMs) across a wide range of tasks. These instructions are often complex, diverse, and subject to frequent change. However, LLMs do not always attend to these instructions reliably, and users lack simple mechanisms to emphasize their importance beyond modifying prompt wording or structure. To address this, we present an inference-time method that enables users to emphasize specific parts of their prompt by steering the model's attention toward them, aligning the model's perceived importance of different prompt tokens with user intent. Unlike prior approaches that are limited to static instructions, require significant offline profiling, or rely on fixed biases, we dynamically update the proportion of model attention given to the user-specified parts--ensuring improved instruction following without performance degradation. We demonstrate that our approach improves instruction following across a variety of tasks involving multiple instructions and generalizes across models of varying scales. |
| title | Spotlight Your Instructions: Instruction-following with Dynamic Attention Steering |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2505.12025 |