Check the problem size, uniformity, resource measure, reduction and approximation factor. A claim in one restricted model should not be promoted to a statement about every algorithm.
Some puzzles seem to get overwhelmingly harder as they grow. A new AI-generated argument claims that a famous prediction about how quickly this happens is wrong.
The determinant is a familiar way to combine a square matrix. Its cousin, the permanent, sums similar products without alternating signs. That small-looking change leads to a very different computational problem.
Complexity theory · provider_announcement
Working largely on its own, Claude designed quantum circuits that improve on the best published versions, in some cases by an order of magnitude or
Automatically screened · not peer reviewed by this siteRead analysis note ↗Complexity theory · provider_announcement
Preliminary interpretability notes on successor heads and on oversampling topics in sparse autoencoder training for safety-relevant features.
Automatically screened · not peer reviewed by this siteRead analysis note ↗Complexity theory · provider_announcement
Preliminary interpretability notes on quantified autointerpretability evaluations for sparse autoencoders and self-explaining features.
Automatically screened · not peer reviewed by this siteRead analysis note ↗Complexity theory · provider_announcement
Five open challenges for mechanistic interpretability, plus notes on linear representations, rare features, and measuring feature sensitivity.
Automatically screened · not peer reviewed by this siteRead analysis note ↗Complexity theory · provider_announcement
Comparing alternative sparse autoencoder architectures on the sparsity and reconstruction tradeoff, plus notes on external interpretability work.
Automatically screened · not peer reviewed by this siteRead analysis note ↗Complexity theory · provider_announcement
Preliminary interpretability notes on scaling laws for dictionary learning, sparse autoencoder training, and circuit-analysis techniques.
Automatically screened · not peer reviewed by this siteRead analysis note ↗Complexity theory · provider_announcement
Developing interpretability ideas on dictionary learning, defining features, and the possibility of superposition in attention heads.
Automatically screened · not peer reviewed by this siteRead analysis note ↗Complexity theory · provider_announcement
Reverse-engineering small attention-only transformers reveals how induction heads form via head composition.
Automatically screened · not peer reviewed by this siteRead analysis note ↗