SAE Interprets MoE: Sparse AutoEncoders Reveal That MoE Experts Exhibit Human-Interpretable Fine-Grained Specialization
cs.LG
Submitted: 2026-11-05
Updated: 2026-11-05
Code: https://github.com/Bytedance-Seed/SAE-Interprets-MoE
Terminology
Sources
- Qwen-Scope: Turning Sparse Features into Development Tools for Large Language Models
- The Llama 3 Herd of Models
- The Expert Strikes Back: Interpreting Mixture-of-Experts Language Models at Expert Level
- Mixtral of Experts
- DeepSeek-V3 Technical Report
- Decoupled Weight Decay Regularization
- Olmo 3
- Automatically Interpreting Millions of Features in Large Language Models
- OpenMoE: An Early Effort on Open Mixture-of-Experts Language Models
- Qwen3 Technical Report
- Memory Retrieval and Consolidation in Large Language Models through Function Tokens
Related papers
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis
- (Mis)generalization of Helpful-Only Fine-tuning
- Provably Learning Diffusion Models under the Manifold Hypothesis: Collapse and Refine