C$^{2}$R: Cross-sample Consistency Regularization Mitigates Feature Splitting and Absorption in Sparse Autoencoders

• URL: http://arxiv.org/abs/2606.30609v1
• Date: Mon, 29 Jun 2026 17:45:31 GMT
• Title: C$^{2}$R: Cross-sample Consistency Regularization Mitigates Feature Splitting and Absorption in Sparse Autoencoders
• Abstract summary: Sparse Autoencoders (SAEs) are widely used to interpret large language models by decomposing activations into sparse, human-understandable features.<n>These issues stem from inconsistent latent assignment across samples.<n>We introduce C$2$R, which penalizes the co-activation of directionally similar latents.
• Score: 56.67585330425431
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Sparse Autoencoders (SAEs) are widely used to interpret large language models by decomposing activations into sparse, human-understandable features, but scaling to large dictionaries exposes fundamental challenges. Systematic studies reveal pervasive feature splitting that fragments coherent concepts into non-atomic latents and widespread feature absorption that creates arbitrary exceptions in general features, severely compromising latent reliability. These issues stem from inconsistent latent assignment across samples: without cross-sample constraints, per-sample optimization often allows a single underlying concept to be inconsistently distributed across multiple redundant or interfering latents. To address this, we introduce C$^2$R (\underline{\textbf{C}}ross-sample \underline{\textbf{C}}onsistency \underline{\textbf{R}}egularization). C$^2$R explicitly encourages that each semantic feature is consistently represented by a unified latent across the batch by penalizing the co-activation of directionally similar latents. Comprehensive evaluation demonstrates that C$^2$R effectively mitigates both splitting and absorption while, crucially, preserving reconstruction fidelity, providing a principled solution that enhances latent interpretability without degrading model performance. Source code is available at https://github.com/hr-jin/Cross-sample-Consistency-Regularization.

Related papers

• Prototype-Anchored Generalized Manifold Regression for Unknown-Domain Object Detection [58.25418970608328]
  We study Single-Domain Generalized Object Detection (Single-DGOD), which aims to transfer a detector trained on a single source domain to multiple unseen domains.<n>Existing methods mainly rely on simulation-driven strategies, such as data augmentation or textual prompts, to enlarge the training distribution.<n>We propose Manifold Regression with Visual-Text Dual Chain-of-Thought (MR-DCoT), which formulates unknown-domain generalization as a manifold regression problem.
  arXiv  Detail & Related papers  (2026-07-08T09:27:15Z)
• Constrained Paraphrase Consistency for LLM Hallucination Detection [11.075787335043238]
  Large language models (LLMs) can generate factually inconsistent claims, motivating accurate and scalable hallucination detectors.<n>We propose Consistency-Constrained Hallucination Detector (CCHD), which formulates training as a constrained optimization problem.
  arXiv  Detail & Related papers  (2026-06-06T13:14:10Z)
• Subspace-Aware Sparse Autoencoders for Effective Mechanistic Interpretability [11.543771846135021]
  Sparse Autoencoders (SAEs) are widely used for mechanistic interpretability in large language models.<n>We show that this assumption mismatches with the multi-dimensional structure of model features.<n>We introduce Subspace-Aware Sparse Autoencoders (SASA), which replace single-vector decoders with learned decoder subspaces.
  arXiv  Detail & Related papers  (2026-06-04T16:08:25Z)
• Feature Starvation as Geometric Instability in Sparse Autoencoders [0.764671395172401]
  We introduce adaptive elastic net SAEs, a fully differentiable architecture grounded in classical sparse regression.<n>AEN-SAEs combine an $ell$ structural term that enforces strong convexity and Lipschitz stability with adaptive $ell$ reweighting.<n>We show that AEN-SAEs yield a Lipschitzcontinuous sparse coding map and recover the global feature support under mild assumptions.
  arXiv  Detail & Related papers  (2026-05-06T18:11:14Z)
• Bi-Lipschitz Autoencoder With Injectivity Guarantee [14.99578789813769]
  Bi-Lipschitz Autoencoder (BLAE) is an injective regularization scheme based on a separation criterion to eliminate pathological local minima.<n>BLAE consistently outperforms existing methods in preserving manifold structure while remaining resilient to sampling sparsity and distribution shifts.
  arXiv  Detail & Related papers  (2026-04-08T05:40:00Z)
• SIM-CoT: Supervised Implicit Chain-of-Thought [108.30049193668083]
  Implicit Chain-of-Thought (CoT) methods offer a token-efficient alternative to explicit CoT reasoning in Large Language Models.<n>We identify a core latent instability issue when scaling the computational budget of implicit CoT.<n>We propose SIM-CoT, a plug-and-play training module that introduces step-level supervision to stabilize and enrich the latent reasoning space.
  arXiv  Detail & Related papers  (2025-09-24T17:01:32Z)
• UniMRSeg: Unified Modality-Relax Segmentation via Hierarchical Self-Supervised Compensation [104.59740403500132]
  Multi-modal image segmentation faces real-world deployment challenges from incomplete/corrupted modalities degrading performance.<n>We propose a unified modality-relax segmentation network (UniMRSeg) through hierarchical self-supervised compensation (HSSC)<n>Our approach hierarchically bridges representation gaps between complete and incomplete modalities across input, feature and output levels.
  arXiv  Detail & Related papers  (2025-09-19T17:29:25Z)
• PCSR: Pseudo-label Consistency-Guided Sample Refinement for Noisy Correspondence Learning [17.302186298424836]
  Cross-modal retrieval aims to align different modalities via semantic similarity.<n>Existing methods often assume that image-text pairs are perfectly aligned, overlooking Noisy Correspondences in real data.
  arXiv  Detail & Related papers  (2025-09-19T05:41:17Z)
• Tackling Ambiguity from Perspective of Uncertainty Inference and Affinity Diversification for Weakly Supervised Semantic Segmentation [12.308473939796945]
  Weakly supervised semantic segmentation (WSSS) with image-level labels aims to achieve dense tasks without laborious annotations. The performance of WSSS, especially the stages of generating Class Activation Maps (CAMs) and refining pseudo masks, widely suffers from ambiguity. We propose UniA, a unified single-staged WSSS framework, to tackle this issue from the perspective of uncertainty inference and affinity diversification.
  arXiv  Detail & Related papers  (2024-04-12T01:54:59Z)
• Improve Variational Autoencoder for Text Generationwith Discrete Latent Bottleneck [52.08901549360262]
  Variational autoencoders (VAEs) are essential tools in end-to-end representation learning. VAEs tend to ignore latent variables with a strong auto-regressive decoder. We propose a principled approach to enforce an implicit latent feature matching in a more compact latent space.
  arXiv  Detail & Related papers  (2020-04-22T14:41:37Z)
• Distributional Robustness and Regularization in Reinforcement Learning [62.23012916708608]
  We introduce a new regularizer for empirical value functions and show that it lower bounds the Wasserstein distributionally robust value function. It suggests using regularization as a practical tool for dealing with $textitexternal uncertainty$ in reinforcement learning.
  arXiv  Detail & Related papers  (2020-03-05T19:56:23Z)

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.