Tag: Interpretability
Qwen-Scope: Alibaba’s Open ‘X-Ray’ for Model Interpretability
Alibaba releases Qwen-Scope, a massive collection of Sparse Autoencoders (SAEs) that allows researchers to 'look inside' Qwen models and steer their behavior.
Assistant Axis: Controlling LLM Character
Anthropic research reveals how LLMs drift between personas. The Assistant Axis stabilizes model behavior and prevents harmful outputs via activation capping.
OpenAI: Interpretability via Sparse Circuits
OpenAI researchers publish findings on Sparse Circuits, revealing how neural networks process information through dedicated, interpretable internal pathways.