会员服务
Behavioral Steering in a 35B MoE Language Model via SAE-Decoded Probe Vectors: One Agency Axis, Not Five Traits
Arxiv
0+阅读 · 3月17日
Capability-Guided Compression: Toward Interpretability-Aware Budget Allocation for Large Language Models
Capability-Guided Compression: Toward Interpretability-Aware Budget Allocation for Large Language Models
Arxiv
0+阅读 · 3月17日
MO-SAE:Multi-Objective Stacked Autoencoders Optimization for Edge Anomaly Detection
Arxiv
0+阅读 · 3月14日
Facet-Level Persona Control by Trait-Activated Routing with Contrastive SAE for Role-Playing LLMs
Arxiv
0+阅读 · 2月22日
Sparse Autoencoders Can Capture Language-Specific Concepts Across Diverse Languages
Arxiv
0+阅读 · 2月23日
Stable and Steerable Sparse Autoencoders with Weight Regularization
Arxiv
0+阅读 · 3月4日
Mechanistic Interpretability with Sparse Autoencoder Neural Operators
Arxiv
0+阅读 · 2月23日
How Do LLMs Encode Scientific Quality? An Empirical Study Using Monosemantic Features from Sparse Autoencoders
Arxiv
0+阅读 · 2月22日
Unveiling Decision-Making in LLMs for Text Classification : Extraction of influential and interpretable concepts with Sparse Autoencoders
Arxiv
0+阅读 · 2月20日
Step-Level Sparse Autoencoder for Reasoning Process Interpretation
Arxiv
0+阅读 · 3月3日
OrthoEraser: Coupled-Neuron Orthogonal Projection for Concept Erasure
Arxiv
0+阅读 · 3月12日
Sparse Autoencoders for Sequential Recommendation Models: Interpretation and Flexible Control
Arxiv
0+阅读 · 2月17日
Toward Faithful Retrieval-Augmented Generation with Sparse Autoencoders
Arxiv
0+阅读 · 2月11日
AudioSAE: Towards Understanding of Audio-Processing Models with Sparse AutoEncoders
Arxiv
0+阅读 · 2月6日
Endogenous Resistance to Activation Steering in Language Models
Arxiv
0+阅读 · 2月6日
参考链接
微信扫码咨询专知VIP会员
Top