The Mechanism of Weak-to-Strong Generalization: Feature Elicitation from Latent Knowledge

• URL: http://arxiv.org/abs/2605.12908v1
• Date: Wed, 13 May 2026 02:35:58 GMT
• Title: The Mechanism of Weak-to-Strong Generalization: Feature Elicitation from Latent Knowledge
• Abstract summary: Weak-to-strong (W2S) generalization is proposed as an approach to aligning AI systems.<n>We study W2S in the setting of reward-model learning with two-layer neural networks.<n>We prove that the strong model efficiently learns task $$, eliciting its pre-trained knowledge while retaining general capabilities.
• Score: 51.56484100374058
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Weak-to-strong (W2S) generalization, in which a strong model is fine-tuned on outputs of a weaker, task-specialized model, has been proposed as an approach to aligning superhuman AI systems. Existing theoretical analyses either fix the student's representations or operate in restricted settings. Whether multi-step SGD can succeed in feature learning while preserving diverse pre-trained capabilities remains open. We study W2S in the setting of reward-model learning with two-layer neural networks. The strong model has pre-trained representations organized into low-dimensional subspaces $V_k$, and is fine-tuned under the supervision of a weak model specialized on task $κ$. We prove that the strong model efficiently learns task $κ$, eliciting its pre-trained knowledge while retaining general capabilities. This establishes W2S generalization in the feature-learning regime, in the sense that the strong model acquires the target feature direction through W2S training, rather than having it given a priori. Moreover, W2S preserves pre-trained off-target features, whereas standard supervised fine-tuning causes catastrophic forgetting when off-target feature directions are correlated with the target's. Numerical experiments on synthetic data confirm our theoretical results.

Related papers

• On the Blessing of Pre-training in Weak-to-Strong Generalization [38.765691061220686]
  We identify pre-training as the prerequisite for the emergence of Weak-to-Strong Generalization (W2SG)<n>We demonstrate that W2SG is not an innate capability but emerges via a phase transition tightly coupled with pre-training.
  arXiv  Detail & Related papers  (2026-05-07T05:55:10Z)
• Why Does Reinforcement Learning Generalize? A Feature-Level Mechanistic Study of Post-Training in Large Language Models [51.45122910254346]
  Reinforcement learning (RL)-based post-training often improves the reasoning performance of large language models beyond the training domain.<n>Supervised fine-tuning (SFT) frequently leads to general forgetting capabilities.<n>We present a feature-level mechanistic analysis methodology to probe RL generalization using a controlled experimental setup.
  arXiv  Detail & Related papers  (2026-04-27T21:22:34Z)
• How LLMs Learn to Reason: A Complex Network Perspective [14.638878448692493]
  Training large language models with Reinforcement Learning from Verifiable Rewards exhibits a set of puzzling behaviors.<n>We propose that these seemingly disparate phenomena can be explained using a single unifying theory.<n>Our work provides a new physical intuition for engineering the emergent reasoning capabilities of future AI systems.
  arXiv  Detail & Related papers  (2025-09-28T04:10:37Z)
• Sycophancy Mitigation Through Reinforcement Learning with Uncertainty-Aware Adaptive Reasoning Trajectories [58.988535279557546]
  We introduce textbf sycophancy Mitigation through Adaptive Reasoning Trajectories.<n>We show that SMART significantly reduces sycophantic behavior while preserving strong performance on out-of-distribution inputs.
  arXiv  Detail & Related papers  (2025-09-20T17:09:14Z)
• Weak-to-Strong Generalization with Failure Trajectories: A Tree-based Approach to Elicit Optimal Policy in Strong Models [26.393123295267642]
  Weak-to-Strong generalization (W2SG) is a new trend to elicit the full capabilities of a strong model with supervision from a weak model.<n>We fine-tune a strong model with trajectories of intermediate actions generated by a weak model.<n>Our empirical evaluations demonstrate substantial improvements in reasoning and decision-making capabilities across diverse task domains.
  arXiv  Detail & Related papers  (2025-07-25T00:17:09Z)
• On Weak-to-Strong Generalization and f-Divergence [23.062111583403095]
  Weak-to-strong generalization (W2SG) has emerged as a promising paradigm for stimulating the capabilities of strong pre-trained models.<n>We introduce $f$-divergence as an information-theoretic loss function framework in W2SG.<n>We empirically demonstrate that $f$-divergence loss, which generalizes widely-used metrics like KL divergence, effectively improves generalization and noise tolerance of the strong model in practice.
  arXiv  Detail & Related papers  (2025-06-03T17:40:08Z)
• LARES: Latent Reasoning for Sequential Recommendation [96.26996622771593]
  We present LARES, a novel and scalable LAtent REasoning framework for Sequential recommendation.<n>Our proposed approach employs a recurrent architecture that allows flexible expansion of reasoning depth without increasing parameter complexity.<n>Our framework exhibits seamless compatibility with existing advanced models, further improving their recommendation performance.
  arXiv  Detail & Related papers  (2025-05-22T16:22:54Z)
• UniErase: Towards Balanced and Precise Unlearning in Language Models [69.04923022755547]
  Large language models (LLMs) require iterative updates to address the outdated information problem.<n>UniErase is a novel unlearning framework that demonstrates precision and balanced performances between knowledge unlearning and ability retaining.
  arXiv  Detail & Related papers  (2025-05-21T15:53:28Z)
• Towards Generalized Video Quality Assessment: A Weak-to-Strong Learning Paradigm [76.63001244080313]
  Video quality assessment (VQA) seeks to predict the perceptual quality of a video in alignment with human visual perception.<n>The dominant VQA paradigm relies on supervised training with human-labeled datasets.<n>We explore weak-to-strong (W2S) learning as a new paradigm for advancing VQA without reliance on large-scale human-labeled datasets.
  arXiv  Detail & Related papers  (2025-05-06T15:29:32Z)
• TWINS: A Fine-Tuning Framework for Improved Transferability of Adversarial Robustness and Generalization [89.54947228958494]
  This paper focuses on the fine-tuning of an adversarially pre-trained model in various classification tasks. We propose a novel statistics-based approach, Two-WIng NormliSation (TWINS) fine-tuning framework. TWINS is shown to be effective on a wide range of image classification datasets in terms of both generalization and robustness.
  arXiv  Detail & Related papers  (2023-03-20T14:12:55Z)

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.