Context Training with Active Information Seeking

• URL: http://arxiv.org/abs/2605.13050v2
• Date: Thu, 14 May 2026 12:47:52 GMT
• Title: Context Training with Active Information Seeking
• Abstract summary: We equip contexts with Wikipedia search and browser tools for active information seeking.<n>We show that when paired with a search-based training procedure that maintains and prunes multiple candidate contexts, active information seeking delivers consistent and substantial gains.
• Score: 23.67605567495757
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Most existing large language models (LLMs) are expensive to adapt after deployment, especially when a task requires newly produced information or niche domain knowledge. Recent work has shown that, by manipulating and optimizing their context, LLMs can be tailored to downstream tasks without updating their weights. However, most existing methods remain closed-loop, relying solely on the model's intrinsic knowledge. In this paper, we equip these context optimizers with Wikipedia search and browser tools for active information seeking. We show that naively adding these tools to a standard sequential context optimization pipeline can actually degrade performance compared to baselines. However, when paired with a search-based training procedure that maintains and prunes multiple candidate contexts, active information seeking delivers consistent and substantial gains. We demonstrate these improvements across diverse domains, including low-resource translation (Flores+), health scenarios (HealthBench), and reasoning-heavy tasks (LiveCodeBench and Humanity's Last Exam). Furthermore, our method proves to be data-efficient, robust across different hyperparameters, and capable of generating effective textual contexts that generalize well across different models.

Related papers

• SIEVE: Sample-Efficient Parametric Learning from Natural Language [66.46967795931559]
  We propose SIEVE, a method for sample-efficient parametric learning from natural language context.<n>We use a novel synthetic data generation pipeline, SIEVE-GEN, that leverages the insight that context is decomposable.<n>Our results show that SIEVE outperforms prior context distillation methods using just three query examples.
  arXiv  Detail & Related papers  (2026-02-02T18:52:52Z)
• URLs Help, Topics Guide: Understanding Metadata Utility in LLM Training [33.68104398807581]
  We show that only URL context speeds up training, whereas quality scores and topic/format domain information offer no clear benefit.<n>Although topic and format metadata do not accelerate training, they are effective for steering outputs, offering human-interpretable control over generation.
  arXiv  Detail & Related papers  (2025-05-22T12:01:20Z)
• Learning Task Representations from In-Context Learning [67.66042137487287]
  Large language models (LLMs) have demonstrated remarkable proficiency in in-context learning (ICL)<n>We introduce an automated formulation for encoding task information in ICL prompts as a function of attention heads.<n>The proposed method successfully extracts task-specific information from in-context demonstrations and excels in both text and regression tasks.
  arXiv  Detail & Related papers  (2025-02-08T00:16:44Z)
• Reference Trustable Decoding: A Training-Free Augmentation Paradigm for Large Language Models [79.41139393080736]
  Large language models (LLMs) have rapidly advanced and demonstrated impressive capabilities. In-Context Learning (ICL) and. Efficient Fine-Tuning (PEFT) are currently two mainstream methods for augmenting. LLMs to downstream tasks. We propose Reference Trustable Decoding (RTD), a paradigm that allows models to quickly adapt to new tasks without fine-tuning.
  arXiv  Detail & Related papers  (2024-09-30T10:48:20Z)
• Prompt Refinement or Fine-tuning? Best Practices for using LLMs in Computational Social Science Tasks [0.0]
  We present an overview of the performance of modern LLM-based classification methods on a benchmark of 23 social knowledge tasks. Our results point to three best practices: select models with larger vocabulary and pre-training corpora; avoid simple zero-shot in favor of AI-enhanced prompting; fine-tune on task-specific data.
  arXiv  Detail & Related papers  (2024-08-02T15:46:36Z)
• Robust and Scalable Model Editing for Large Language Models [75.95623066605259]
  We propose EREN (Edit models by REading Notes) to improve the scalability and robustness of LLM editing. Unlike existing techniques, it can integrate knowledge from multiple edits, and correctly respond to syntactically similar but semantically unrelated inputs.
  arXiv  Detail & Related papers  (2024-03-26T06:57:23Z)
• Learning to Reduce: Optimal Representations of Structured Data in Prompting Large Language Models [42.16047343029512]
  Large Language Models (LLMs) have been widely used as general-purpose AI agents. We propose a framework, Learning to Reduce, that fine-tunes a language model to generate a reduced version of an input context. We show that our model achieves comparable accuracies in selecting the relevant evidence from an input context.
  arXiv  Detail & Related papers  (2024-02-22T00:41:23Z)
• Vocabulary-Defined Semantics: Latent Space Clustering for Improving In-Context Learning [32.178931149612644]
  In-context learning enables language models to adapt to downstream data or incorporate tasks by few samples as demonstrations within the prompts. However, the performance of in-context learning can be unstable depending on the quality, format, or order of demonstrations. We propose a novel approach "vocabulary-defined semantics"
  arXiv  Detail & Related papers  (2024-01-29T14:29:48Z)
• Structured Packing in LLM Training Improves Long Context Utilization [11.484631908171465]
  This study investigates structuring training data to enhance semantic interdependence.<n>We introduce the Structured Packing for Long Context (SPLiCe) method.<n>We validate SPLiCe empirically across models of varying sizes.
  arXiv  Detail & Related papers  (2023-12-28T16:25:52Z)
• Pre-Training to Learn in Context [138.0745138788142]
  The ability of in-context learning is not fully exploited because language models are not explicitly trained to learn in context. We propose PICL (Pre-training for In-Context Learning), a framework to enhance the language models' in-context learning ability. Our experiments show that PICL is more effective and task-generalizable than a range of baselines, outperforming larger language models with nearly 4x parameters.
  arXiv  Detail & Related papers  (2023-05-16T03:38:06Z)
• Efficient Nearest Neighbor Language Models [114.40866461741795]
  Non-parametric neural language models (NLMs) learn predictive distributions of text utilizing an external datastore. We show how to achieve up to a 6x speed-up in inference speed while retaining comparable performance.
  arXiv  Detail & Related papers  (2021-09-09T12:32:28Z)
• Pre-training Text Representations as Meta Learning [113.3361289756749]
  We introduce a learning algorithm which directly optimize model's ability to learn text representations for effective learning of downstream tasks. We show that there is an intrinsic connection between multi-task pre-training and model-agnostic meta-learning with a sequence of meta-train steps.
  arXiv  Detail & Related papers  (2020-04-12T09:05:47Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.