Knowledge Collapse in LLMs: When Fluency Survives but Facts Fail under Recursive Synthetic Training

• URL: http://arxiv.org/abs/2509.04796v1
• Date: Fri, 05 Sep 2025 04:29:15 GMT
• Title: Knowledge Collapse in LLMs: When Fluency Survives but Facts Fail under Recursive Synthetic Training
• Authors: Figarri Keisha, Zekun Wu, Ze Wang, Adriano Koshiyama, Philip Treleaven,
• Abstract summary: Large language models increasingly rely on synthetic data due to human-written content scarcity.<n>Recursive training on model-generated outputs leads to model collapse, a degenerative process threatening factual reliability.
• Score: 2.094557609248011
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Large language models increasingly rely on synthetic data due to human-written content scarcity, yet recursive training on model-generated outputs leads to model collapse, a degenerative process threatening factual reliability. We define knowledge collapse as a distinct three-stage phenomenon where factual accuracy deteriorates while surface fluency persists, creating "confidently wrong" outputs that pose critical risks in accuracy-dependent domains. Through controlled experiments with recursive synthetic training, we demonstrate that collapse trajectory and timing depend critically on instruction format, distinguishing instruction-following collapse from traditional model collapse through its conditional, prompt-dependent nature. We propose domain-specific synthetic training as a targeted mitigation strategy that achieves substantial improvements in collapse resistance while maintaining computational efficiency. Our evaluation framework combines model-centric indicators with task-centric metrics to detect distinct degradation phases, enabling reproducible assessment of epistemic deterioration across different language models. These findings provide both theoretical insights into collapse dynamics and practical guidance for sustainable AI training in knowledge-intensive applications where accuracy is paramount.

Related papers

• Human Supervision as an Information Bottleneck: A Unified Theory of Error Floors in Human-Guided Learning [51.56484100374058]
  We argue that limitations reflect structural properties of the supervision channel rather than model scale or optimization.<n>We develop a unified theory showing that whenever the human supervision channel is not sufficient for a latent evaluation target, it acts as an information-reducing channel.
  arXiv  Detail & Related papers  (2026-02-26T19:11:32Z)
• Bridging Information Asymmetry: A Hierarchical Framework for Deterministic Blind Face Restoration [31.878334664450776]
  We present textbfPrefRestore, a hierarchical framework that integrates discrete semantic logic with continuous texture generation.<n>Our methodology fundamentally addresses this information disparity through two complementary strategies.<n>Pref-Restore achieves state-of-the-art performance across synthetic and real-world benchmarks.
  arXiv  Detail & Related papers  (2026-01-27T11:50:31Z)
• On the Limits of Self-Improving in LLMs and Why AGI, ASI and the Singularity Are Not Near Without Symbolic Model Synthesis [0.01269104766024433]
  We formalise self-training in Large Language Models (LLMs) and Generative AI as a discrete-time dynamical system.<n>We derive two fundamental failure modes: (1) Entropy Decay, where finite sampling effects cause a monotonic loss of distributional diversity (mode collapse), and (2) Variance Amplification, where the loss of external grounding causes the model's representation of truth to drift as a random walk.
  arXiv  Detail & Related papers  (2026-01-05T19:50:49Z)
• Provable Benefit of Curriculum in Transformer Tree-Reasoning Post-Training [76.12556589212666]
  We show that curriculum post-training avoids the exponential complexity bottleneck.<n>Under outcome-only reward signals, reinforcement learning finetuning achieves high accuracy with sample complexity.<n>We establish guarantees for test-time scaling, where curriculum-aware querying reduces both reward oracle calls and sampling cost from exponential to order.
  arXiv  Detail & Related papers  (2025-11-10T18:29:54Z)
• ForTIFAI: Fending Off Recursive Training Induced Failure for AI Model Collapse [13.096745830570944]
  We introduce the Truncated-Cross-Entropy (TCE) loss function to mitigate model collapse in synthetic data.<n>TCE mitigates collapse by selectively ignoring high-confidence tokens during training, effectively filtering out likely machine-generated artifacts.<n>Our results demonstrate that confidence-aware training objectives can substantially delay collapse onset, offering a practical and generalizable tool for model robustness under synthetic-data exposure.
  arXiv  Detail & Related papers  (2025-09-10T20:06:51Z)
• A Theoretical Perspective: How to Prevent Model Collapse in Self-consuming Training Loops [55.07063067759609]
  High-quality data is essential for training large generative models, yet the vast reservoir of real data available online has become nearly depleted.<n>Models increasingly generate their own data for further training, forming Self-consuming Training Loops (STLs)<n>Some models degrade or even collapse, while others successfully avoid these failures, leaving a significant gap in theoretical understanding.
  arXiv  Detail & Related papers  (2025-02-26T06:18:13Z)
• Causality can systematically address the monsters under the bench(marks) [64.36592889550431]
  Benchmarks are plagued by various biases, artifacts, or leakage.<n>Models may behave unreliably due to poorly explored failure modes.<n> causality offers an ideal framework to systematically address these challenges.
  arXiv  Detail & Related papers  (2025-02-07T17:01:37Z)
• Bridging Interpretability and Robustness Using LIME-Guided Model Refinement [0.0]
  Local Interpretable Model-Agnostic Explanations (LIME) systematically enhance model robustness.<n> Empirical evaluations on multiple benchmark datasets demonstrate that LIME-guided refinement not only improves interpretability but also significantly enhances resistance to adversarial perturbations and generalization to out-of-distribution data.
  arXiv  Detail & Related papers  (2024-12-25T17:32:45Z)
• Unveiling and Addressing Pseudo Forgetting in Large Language Models [17.888328120571245]
  We show that the performance degradation on previous tasks is not attributed to a loss of capabilities, but rather to the failure of the instructions to activate the appropriate model abilities.<n>We propose Rationale-Guidance Difficulty based Replay (RGD-R) framework that dynamically allocates replay data based on the model's ability to correctly leverage the intrinsic capabilities.
  arXiv  Detail & Related papers  (2024-11-18T14:28:04Z)
• Learning by Surprise: Surplexity for Mitigating Model Collapse in Generative AI [1.6545633988217645]
  As synthetic content infiltrates the web, generative AI models may be retrained on their own outputs.<n>This leads to model collapse: a progressive loss of performance and diversity across generations.<n>We introduce new measures that characterise collapse directly from a model's next-token probability distributions.
  arXiv  Detail & Related papers  (2024-10-16T08:02:48Z)
• Interpretable Imitation Learning with Dynamic Causal Relations [65.18456572421702]
  We propose to expose captured knowledge in the form of a directed acyclic causal graph. We also design this causal discovery process to be state-dependent, enabling it to model the dynamics in latent causal graphs. The proposed framework is composed of three parts: a dynamic causal discovery module, a causality encoding module, and a prediction module, and is trained in an end-to-end manner.
  arXiv  Detail & Related papers  (2023-09-30T20:59:42Z)
• An Experimental Study of Semantic Continuity for Deep Learning Models [11.883949320223078]
  We argue that semantic discontinuity results from inappropriate training targets and contributes to notorious issues such as adversarial robustness, interpretability, etc. We first conduct data analysis to provide evidence of semantic discontinuity in existing deep learning models, and then design a simple semantic continuity constraint which theoretically enables models to obtain smooth gradients and learn semantic-oriented features.
  arXiv  Detail & Related papers  (2020-11-19T12:23:28Z)
• Gradient Starvation: A Learning Proclivity in Neural Networks [97.02382916372594]
  Gradient Starvation arises when cross-entropy loss is minimized by capturing only a subset of features relevant for the task. This work provides a theoretical explanation for the emergence of such feature imbalance in neural networks.
  arXiv  Detail & Related papers  (2020-11-18T18:52:08Z)

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.