Context-Informed Grounding Supervision

• URL: http://arxiv.org/abs/2506.15480v1
• Date: Wed, 18 Jun 2025 14:13:56 GMT
• Title: Context-Informed Grounding Supervision
• Authors: Hyunji Lee, Seunghyun Yoon, Yunjae Won, Hanseok Oh, Geewook Kim, Trung Bui, Franck Dernoncourt, Elias Stengel-Eskin, Mohit Bansal, Minjoon Seo,
• Abstract summary: Context-INformed Grounding Supervision (CINGS) is a post-training supervision in which the model is trained with relevant context prepended to the response.<n>Our experiments demonstrate that models trained with CINGS exhibit stronger grounding in both textual and visual domains.
• Score: 102.11698329887226
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models (LLMs) are often supplemented with external knowledge to provide information not encoded in their parameters or to reduce hallucination. In such cases, we expect the model to generate responses by grounding its response in the provided external context. However, prior work has shown that simply appending context at inference time does not ensure grounded generation. To address this, we propose Context-INformed Grounding Supervision (CINGS), a post-training supervision in which the model is trained with relevant context prepended to the response, while computing the loss only over the response tokens and masking out the context. Our experiments demonstrate that models trained with CINGS exhibit stronger grounding in both textual and visual domains compared to standard instruction-tuned models. In the text domain, CINGS outperforms other training methods across 11 information-seeking datasets and is complementary to inference-time grounding techniques. In the vision-language domain, replacing a vision-language model's LLM backbone with a CINGS-trained model reduces hallucinations across four benchmarks and maintains factual consistency throughout the generated response. This improved grounding comes without degradation in general downstream performance. Finally, we analyze the mechanism underlying the enhanced grounding in CINGS and find that it induces a shift in the model's prior knowledge and behavior, implicitly encouraging greater reliance on the external context.

Related papers

• "Lost-in-the-Later": Framework for Quantifying Contextual Grounding in Large Language Models [4.712325494028972]
  We introduce CoPE, a novel evaluation framework that measures contextual knowledge across models and languages.<n>We analyze how large language models integrate context, prioritize information, and incorporate PK in open-ended question answering.<n>We find that reasoning models, as well as non-reasoning models prompted with chain-of-thought (CoT), use context even less than non-reasoning models without CoT and fail to mitigate the lost-in-the-later effect.
  arXiv  Detail & Related papers  (2025-07-07T19:13:20Z)
• SCOPE: A Self-supervised Framework for Improving Faithfulness in Conditional Text Generation [55.61004653386632]
  Large Language Models (LLMs) often produce hallucinations, i.e., information that is unfaithful or not grounded in the input context.<n>This paper introduces a novel self-supervised method for generating a training set of unfaithful samples.<n>We then refine the model using a training process that encourages the generation of grounded outputs over unfaithful ones.
  arXiv  Detail & Related papers  (2025-02-19T12:31:58Z)
• On the Loss of Context-awareness in General Instruction Fine-tuning [101.03941308894191]
  We investigate the loss of context awareness after supervised fine-tuning.<n>We find that the performance decline is associated with a bias toward different roles learned during conversational instruction fine-tuning.<n>We propose a metric to identify context-dependent examples from general instruction fine-tuning datasets.
  arXiv  Detail & Related papers  (2024-11-05T00:16:01Z)
• Grounded Image Text Matching with Mismatched Relation Reasoning [39.524420144738684]
  Grounded Image Text Matching with Mismatched Relation (GITM-MR) is a novel visual-linguistic joint task. GITM-MR requires a model to first determine if an expression describes an image, then localize referred objects or ground the mismatched parts of the text. We propose the Relation-sensitive Correspondence Reasoning Network (RCRN), which incorporates relation-aware reasoning via bi-directional message propagation.
  arXiv  Detail & Related papers  (2023-08-02T15:44:36Z)
• Large Language Models with Controllable Working Memory [64.71038763708161]
  Large language models (LLMs) have led to a series of breakthroughs in natural language processing (NLP) What further sets these models apart is the massive amounts of world knowledge they internalize during pretraining. How the model's world knowledge interacts with the factual information presented in the context remains under explored.
  arXiv  Detail & Related papers  (2022-11-09T18:58:29Z)
• How Context Affects Language Models' Factual Predictions [134.29166998377187]
  We integrate information from a retrieval system with a pre-trained language model in a purely unsupervised way. We report that augmenting pre-trained language models in this way dramatically improves performance and that the resulting system, despite being unsupervised, is competitive with a supervised machine reading baseline.
  arXiv  Detail & Related papers  (2020-05-10T09:28:12Z)
• A Controllable Model of Grounded Response Generation [122.7121624884747]
  Current end-to-end neural conversation models inherently lack the flexibility to impose semantic control in the response generation process. We propose a framework that we call controllable grounded response generation (CGRG) We show that using this framework, a transformer based model with a novel inductive attention mechanism, trained on a conversation-like Reddit dataset, outperforms strong generation baselines.
  arXiv  Detail & Related papers  (2020-05-01T21:22:08Z)

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.