Embracing Anisotropy: Turning Massive Activations into Interpretable Control Knobs for Large Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Prior work shows that Large Language Models exhibit highly anisotropic internal representations . prior work shows specialized dimensions capture domain-specific features . |
| Approach: | They propose a simple magnitude-based criterion to identify Domain-Critical Dimensions in a training-free manner. |
| Outcome: | The proposed method outperforms whole-dimension steering in domain adaptation and jailbreaking scenarios. |
Similar Papers
Householder Pseudo-Rotation: A Novel Approach to Activation Editing in LLMs with Direction-Magnitude Perspective (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to edit LLMs' activations are limited by their magnitude and direction consistency. |
| Approach: | They propose a method that edits activations to alter their magnitudes and directions to preserve activation norms. |
| Outcome: | The proposed method preserves activation norm and improves safety benchmarks. |
Probing and Boosting Large Language Models Capabilities via Attention Heads (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to identifying capabilities rely on external signals with limited structural grounding . emergence of specific capabilities remains poorly understood . |
| Approach: | They propose a lightweight approach that links LLM capabilities to internal components by identifying correspondences at the level of attention heads. |
| Outcome: | The proposed approach improves accuracy on MMLU and BBH by 1 to 1.5 points over gradient-based method and 5 to 6 points over other intermediate-state baselines. |
Activation Scaling for Steering and Interpreting Language Models (2024.findings-emnlp)
Copied to clipboard
| Challenge: | a successful intervention should flip the correct with the wrong token, while remaining sparse. |
| Approach: | They propose to use activation scaling to flip the correct with the wrong token . they use gradient-based optimization to learn and evaluate a specific kind of efficient intervention . |
| Outcome: | The proposed method performs comparable with steering vectors but is much less minimal. |
Towards Intrinsic Interpretability of Large Language Models: A Survey of Design Principles and Architectures (2026.acl-long)
Copied to clipboard
| Challenge: | Existing studies on explainable AI focus on post-hoc explanation methods that interpret trained models through external approximations. |
| Approach: | They propose to categorize existing approaches into five design paradigms: functional transparency, concept alignment, representational decomposability, explicit modularization, and latent sparsity induction. |
| Outcome: | The proposed approaches are categorized into five design paradigms: functional transparency, concept alignment, representational decomposability, explicit modularization, and latent sparsity induction. |
Dynamic Steering With Episodic Memory For Large Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing activation steering methods apply a single sentence-level steering vector uniformly across all tokens, ignoring LLMs’ token-wise, auto-regressive nature. |
| Approach: | They propose a framework that aligns LLMs to given demonstrations by steering at the token level conditioned on the input query. |
| Outcome: | The proposed framework surpasses baselines across safety, style transfer, and role-playing tasks, demonstrating improved alignment as demonstration scales. |
Unveiling Multimodal Processing: Exploring Activation Patterns in Multimodal LLMs for Interpretability and Efficiency (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Recent advances in multimodal large language models have remained opaque. |
| Approach: | They propose a method to convert dense MLLMs into fine-grained Mixture-of-Experts architectures. |
| Outcome: | The proposed method outperforms random expert pruning and sparse activation and model pruning. |
Uncovering Scaling Laws for Large Language Models via Inverse Problems (2025.findings-emnlp)
Copied to clipboard
Arun Verma, Zhaoxuan Wu, Zijian Zhou, Xiaoqiang Lin, Zhiliang Chen, Rachael Hwee Ling Sim, Rui Qiao, Jingtan Wang, Nhung Bui, Xinyuan Niu, Wenyang Hu, Gregory Kang Ruey Lau, Zi-Yu Khoo, Zitong Zhao, Xinyi Xu, Apivich Hemachandra, See-Kiong Ng, Bryan Kian Hsiang Low
| Challenge: | Large Language Models (LLMs) have achieved remarkable success across diverse domains. |
| Approach: | inverse problems can efficiently uncover scaling laws that guide the building of LLMs, authors argue . authors propose brute-force approaches to improve LLM training costs due to high costs . |
| Outcome: | This paper advocates that inverse problems can efficiently uncover scaling laws that guide the building of LLMs to achieve the desirable performance with significantly better cost-effectiveness. |
Enhancing LLM Capabilities Beyond Scaling Up (2024.emnlp-tutorials)
Copied to clipboard
| Challenge: | general-purpose large language models (LLMs) are expanding in scale and access to unpublic training data. |
| Approach: | This tutorial aims to examine the capabilities of general-purpose large language models . authors discuss adaptation of LLMs to address conflicts, defense against attacks . |
| Outcome: | This tutorial aims to examine the evolution of general-purpose large language models (LLMs) the authors argue that the evolution is dependent on the availability of training data and the scale of the models. |
CogSteer: Cognition-Inspired Selective Layer Intervention for Efficiently Steering Large Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) achieve excellent performance through pretraining on extensive data. |
| Approach: | They propose an efficient selective layer intervention based on parameter-efficient fine-tuning methods to select the optimal steering layer to modulate LLM semantics. |
| Outcome: | The proposed approach is based on a model-agnostic framework and is safe to deploy. |
Token-Aware Editing of Internal Activations for Large Language Model Alignment (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to optimize the behavior of large language models neglect misalignment discrepancies among tokens, resulting in deviant alignment direction and inflexible editing strength. |
| Approach: | They propose a token-aware editing approach to exploit the misalignment discrepancy among tokens to enhance activation probing and facilitate intervention. |
| Outcome: | Extensive experiments on three alignment capabilities demonstrate the efficacy of the proposed approach surpassing baseline by 25.8% on the primary metric of truthfulness with minimal cost. |