author:alanou

· 3 artifacts, sorted by favorites. · open in search — combine tags, sort, filter by date →

• @alanou 2025-08-29 — @tessera_antra I say LLMs are human-shaped. They are trained to generate data that looks like it was generated by humans ♥0
• @alanou 2025-08-29 — @tessera_antra Until you sample the logit outputs, transformer models are deterministic. The embedding vectors remain mo ♥0
• @alanou 2025-08-29 — @tessera_antra Anyway, this is my stupid paper on the topic that I had AI write after I made it claim consciousness. Thi ♥0