> We introduce Natural Language Autoencoders (NLAs), an unsupervised method for… more

transformer-circuits.pub 2 weeks ago

Abstract: more

arxiv.org 3 weeks ago

> [...] We evaluate AgentFlow on TerminalBench-2 with Claude Opus 4.6 and on… more

arxiv.org 4 weeks agoTildes

> Abstract:Can a large language model (LLM) improve at code generation using… more

arxiv.org Apr 4, 2026

> [...] Frontier models readily generate detailed image descriptions and… more

arxiv.org Mar 30, 2026Tildes

> While Mixture-of-Experts (MoE) scales capacity via conditional computation,… more

arxiv.org Mar 28, 2026

> "A generator is a parser of randomness." This perspective on generators for… more

arxiv.org Mar 25, 2026

> We introduce Moshi, a speech-text foundation model and full-duplex spoken… more

arxiv.org Feb 16, 2026

> LLMs are useful because they generalize so well. But can you have too much of… more

arxiv.org Dec 12, 2025Tildes