analiza aiqbench
This research introduces HiLight, a framework that enhances frozen LLMs by using reinforcement learning to highlight critical evidence within long contexts without altering the original text or requiring model fine-tuning. By decoupling evidence selection from reasoning, the method improves performance in tasks like sequential recommendation and question answering while demonstrating zero-shot transferability across different solver architectures. This approach addresses a key limitation in current models where valuable information is often lost during context compression or prompt engineering.
Learning Evidence Highlighting for Frozen LLMs
arXiv:2604.22565v3 Announce Type: replace Abstract: Large Language Models (LLMs) can reason well, yet often miss decisive evidence when it is buried in long, noisy contexts. We introduce HiLight, an Evidence Emphasis framework that decouples evidence selection from reasoning for frozen LLM solvers. HiLight avoids compressing or rewriting the input, which can discard or distort evidence, by training a lightweight Emphasis Actor to insert minimal highlight tags around pivotal spans in the unaltered context. A frozen Solver then performs downstream reasoning on the emphasized input. We cast highlighting as a weakly supervised decision-making problem and optimize the Actor with reinforcement learning using only the Solver's task reward, requiring no evidence labels and no access to or modification of the Solver. Across sequential recommendation and long-context question answering, HiLight consistently improves performance over strong prompt-based and automated prompt-optimization baselines. The learned emphasis policy transfers zero-shot to both smaller and larger unseen Solver families, including an API-based Solver, suggesting that the Actor captures genuine, reusable evidence structure rather than overfitting to a single backbone.
Abstract:Large Language Models (LLMs) can reason well, yet often miss decisive evidence when it is buried in long, noisy contexts. We introduce HiLight, an Evidence Emphasis framework that decouples evidence selection from reasoning for frozen LLM solvers. HiLight avoids compressing or rewriting the input, which can discard or distort evidence, by training a lightweight Emphasis Actor to insert minimal highlight tags around pivotal spans in the unaltered context. A frozen Solver then performs downstream reasoning on the emphasized input. We cast highlighting as a weakly supervised decision-making problem and optimize the Actor with reinforcement learning using only the Solver's task reward, requiring no evidence labels and no access to or modification of the Solver. Across sequential recommendation and long-context question answering, HiLight consistently improves performance over strong prompt-based and automated prompt-optimization baselines. The learned emphasis policy transfers zero-shot to both smaller and larger unseen Solver families, including an API-based Solver, suggesting that the Actor captures genuine, reusable evidence structure rather than overfitting to a single backbone.
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
arXiv:2604.22565 [cs.CL]
(or arXiv:2604.22565v3 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2604.22565 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 262 cuvinte
Subiecte conexe
Learning Evidence Highlighting for Frozen LLMs
arXiv · 10 iun. 2026
The user in the LLM might be a moral patient too (a strange implication of the PSM and simulators)
LessWrong · 15 oct. 2026
ReLMem: Learning Recurrent Memory for Longitudinal EHR Modeling
arXiv · 30 sept. 2026
ObjectStream: Latent Objects as Memory Anchors for Streaming Video Understanding
arXiv · 5 aug. 2026