A unifying framework from neural superposition to sparse interpretable codes

References von Helmholtz, H. in The Selected Writings of Hermann von Helmholtz (ed. Karl, R.) (Wesleyan Univ. Press, 1971). McCulloch, W. S. & Pitts, W. A logical calculus of the ideas immanent in nervous activity. Bull. Math. Biophys. 5, 115–133 (1943). Article MathSciNet Google Scholar Kleene, S. in Automata Studies (eds Shannon, C. E. & McCarthy, J.) 3–41 (Princeton Univ. Press, 1956). Ro

References von Helmholtz, H. in The Selected Writings of Hermann von Helmholtz (ed. Karl, R.) (Wesleyan Univ. Press, 1971). McCulloch, W. S. & Pitts, W. A logical calculus of the ideas immanent in nervous activity. Bull. Math. Biophys. 5, 115–133 (1943). Article MathSciNet Google Scholar Kleene, S. in Automata Studies (eds Shannon, C. E. & McCarthy, J.) 3–41 (Princeton Univ. Press, 1956). Rosenblatt, F. The perceptron: a probabilistic model for information storage and organization in the brain. Psychol. Rev. 65, 386–408 (1958). Article Google Scholar Smolensky, P. Tensor product…

related reading

• Toy Models of Superpositiontransformer-circuits.pub
• Towards Monosemanticity: Decomposing Language Models With Dictionary Learningtransformer-circuits.pub
• Verbalizable Representations Form a Global Workspace in Language Modelstransformer-circuits.pub
• The Building Blocks of Interpretabilitydistill.pub
• Aman's AI Journal • Primers • Ilya Sutskever's Top 30aman.ai
• Towards Monosemanticity: Decomposing Language Models With Dictionary Learningtransformer-circuits.pub
• Sparse Autoencoders Find Highly Interpretable Features in Language Modelsarxiv.org
• Language Model Circuits Are Sparse in the Neuron Basis | Transluce AItransluce.org
• Mechanistic Interpretability: A Challenge Common to Both Artificial and Biological Intelligencekempnerinstitute.harvard.edu
• Multimodal Neurons in Artificial Neural Networksdistill.pub
• [2309.08600] Sparse Autoencoders Find Highly Interpretable Features in Language Modelsarxiv.org
• Fact Finding: Attempting to Reverse-Engineer Factual Recall on the Neuron Level (Post 1) — AI Alignment Forumalignmentforum.org