Natural Language Autoencoders \ Anthropic

Turning Claude's thoughts into text

Interpretability Natural Language Autoencoders: Turning Claude’s thoughts into text May 7, 2026 Read the paper When you talk to an AI model like Claude, you talk to it in words. Internally, Claude processes those words as long lists of numbers, before again producing words as its output. These numbers in the middle are called activations— and like neural activity in the human brain, they encode Claude’s thoughts. Also like neural activity, activations are difficult to understand. We can’t easily decode them to read Claude’s thoughts. Over the past few years, we’ve developed a range of tools (l

saved by

• Falah Rajput
• Elizabeth Qiu
• Rikard Saqe
• Lydia Nottingham
• Sarah
• [David L]
• Jo J.
• Kaustubh Kislay
• Lucas Chen

related reading

• Natural Language Autoencoders Produce Unsupervised Explanations of LLM Activationstransformer-circuits.pub
• Tracing the thoughts of a large language model \ Anthropicanthropic.com
• Natural Language Autoencoders Produce Unsupervised Explanations of LLM Activations — LessWronglesswrong.com
• On the Biology of a Large Language Modeltransformer-circuits.pub
• Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnettransformer-circuits.pub
• Transformer Circuits Threadtransformer-circuits.pub
• A global workspace in language models \ Anthropicanthropic.com
• Neuronpedianeuronpedia.org
• GitHub - kitft/natural_language_autoencoders · GitHubgithub.com
• Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnettransformer-circuits.pub
• Emotion concepts and their function in a large language model \ Anthropicanthropic.com
• Mapping the Mind of a Large Language Model \ Anthropicanthropic.com