XPERT: Expert Knowledge Transfer for Effective Training of Language Models

• URL: http://arxiv.org/abs/2605.08842v1
• Date: Sat, 09 May 2026 09:53:03 GMT
• Title: XPERT: Expert Knowledge Transfer for Effective Training of Language Models
• Abstract summary: Mixture-of-Experts (MoE) language models organize knowledge into explicitly routed expert modules.<n>We find that a subset of experts is consistently activated across diverse knowledge domains.<n>We propose XPERT, a framework that extracts, consolidates, and reuses expert knowledge.
• Score: 33.508312430754906
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Mixture-of-Experts (MoE) language models organize knowledge into explicitly routed expert modules, making expert-level representations traceable and analyzable. By analyzing expert activation patterns in MoE large language models (LLMs), we find that a subset of experts is consistently activated across diverse knowledge domains. These common experts encode cross-domain, generalizable knowledge that is closely related to model generalization, naturally raising the question of how such identifiable expert knowledge can be practically reused. Motivated by this observation, we propose XPERT, a framework that extracts, consolidates, and reuses expert knowledge from pre-trained MoE LLMs to support more effective training of language models across different model scales. XPERT identifies cross-domain experts via inference-only analysis, refines their representations through tensor decomposition, and adapts the extracted knowledge to reuse in downstream models. Experiments on language understanding and dialogue generation benchmarks show that models benefiting from reused expert knowledge achieve consistently stronger performance and faster convergence compared to strong baselines. These results highlight MoE LLMs as structured and reusable knowledge sources, and demonstrate the value of expert-level knowledge reuse for improving model training.

Related papers

• Expert Divergence Learning for MoE-based Language Models [21.857507958048696]
  We introduce Expert Divergence Learning, a novel pre-training strategy that explicitly encourages functional specialization among experts.<n>Our method incorporates a label-driven auxiliary loss that leverages domain labels inherent in pre-training corpora.<n>We validate our approach by pre-training MoE models of up to 15 billion parameters from scratch.
  arXiv  Detail & Related papers  (2026-02-10T06:58:38Z)
• Probing the Knowledge Boundary: An Interactive Agentic Framework for Deep Knowledge Extraction [29.717986496967978]
  We propose an interactive agentic framework to systematically extract and quantify the knowledge of Large Language Models.<n>Our method includes four adaptive exploration policies to probe knowledge at different granularities.<n>We observe a clear knowledge scaling law, where larger models consistently extract more knowledge.
  arXiv  Detail & Related papers  (2026-02-01T01:43:44Z)
• The Rise of Parameter Specialization for Knowledge Storage in Large Language Models [50.91855620712756]
  We show that as language models become more advanced, their parameters exhibit increased specialization.<n>We experimentally validate that this specialized distribution of knowledge contributes to improving the efficiency of knowledge utilization in these models.
  arXiv  Detail & Related papers  (2025-05-22T20:15:01Z)
• Large Language Models are Limited in Out-of-Context Knowledge Reasoning [65.72847298578071]
  Large Language Models (LLMs) possess extensive knowledge and strong capabilities in performing in-context reasoning. This paper focuses on a significant aspect of out-of-context reasoning: Out-of-Context Knowledge Reasoning (OCKR), which is to combine multiple knowledge to infer new knowledge.
  arXiv  Detail & Related papers  (2024-06-11T15:58:59Z)
• Translating Expert Intuition into Quantifiable Features: Encode Investigator Domain Knowledge via LLM for Enhanced Predictive Analytics [2.330270848695646]
  This paper explores the potential of Large Language Models to bridge the gap by systematically converting investigator-derived insights into quantifiable, actionable features. We present a framework that leverages LLMs' natural language understanding capabilities to encode these red flags into a structured feature set that can be readily integrated into existing predictive models. The results indicate significant improvements in risk assessment and decision-making accuracy, highlighting the value of blending human experiential knowledge with advanced machine learning techniques.
  arXiv  Detail & Related papers  (2024-05-11T13:23:43Z)
• Commonsense Knowledge Transfer for Pre-trained Language Models [83.01121484432801]
  We introduce commonsense knowledge transfer, a framework to transfer the commonsense knowledge stored in a neural commonsense knowledge model to a general-purpose pre-trained language model. It first exploits general texts to form queries for extracting commonsense knowledge from the neural commonsense knowledge model. It then refines the language model with two self-supervised objectives: commonsense mask infilling and commonsense relation prediction.
  arXiv  Detail & Related papers  (2023-06-04T15:44:51Z)
• Knowledge Rumination for Pre-trained Language Models [77.55888291165462]
  We propose a new paradigm dubbed Knowledge Rumination to help the pre-trained language model utilize related latent knowledge without retrieving it from the external corpus. We apply the proposed knowledge rumination to various language models, including RoBERTa, DeBERTa, and GPT-3.
  arXiv  Detail & Related papers  (2023-05-15T15:47:09Z)
• UNTER: A Unified Knowledge Interface for Enhancing Pre-trained Language Models [100.4659557650775]
  We propose a UNified knowledge inTERface, UNTER, to provide a unified perspective to exploit both structured knowledge and unstructured knowledge. With both forms of knowledge injected, UNTER gains continuous improvements on a series of knowledge-driven NLP tasks.
  arXiv  Detail & Related papers  (2023-05-02T17:33:28Z)
• LM-CORE: Language Models with Contextually Relevant External Knowledge [13.451001884972033]
  We argue that storing large amounts of knowledge in the model parameters is sub-optimal given the ever-growing amounts of knowledge and resource requirements. We present LM-CORE -- a general framework to achieve this -- that allows textitdecoupling of the language model training from the external knowledge source. Experimental results show that LM-CORE, having access to external knowledge, achieves significant and robust outperformance over state-of-the-art knowledge-enhanced language models on knowledge probing tasks.
  arXiv  Detail & Related papers  (2022-08-12T18:59:37Z)
• Generated Knowledge Prompting for Commonsense Reasoning [53.88983683513114]
  We propose generating knowledge statements directly from a language model with a generic prompt format. This approach improves performance of both off-the-shelf and finetuned language models on four commonsense reasoning tasks. Notably, we find that a model's predictions can improve when using its own generated knowledge.
  arXiv  Detail & Related papers  (2021-10-15T21:58:03Z)

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.