Leello
leello.bsky.social
Leello
@leello.bsky.social
CS PhD@EPFL
saturdaygenfo.github.io
The problem is in wrongly understanding the statistics or is it in the belief the question is answerable with an RCT, that anything true is measurable through an average treatment effect?
August 19, 2025 at 5:19 PM
language diffusion models shouldn't require as many neural function calls as generated tokens, saturdaygenfo.github.io/posts/lladap...
trivial improvements on the re-masking strategy of llada
trivial improvements on the re-masking strategy of llada an auto-regressive (AR) language model generates a single token per function call or neural function evaluation (NFE). Language diffusion model...
saturdaygenfo.github.io
July 11, 2025 at 3:33 PM
Reposted by Leello
nihilist Kolmogorovism (probability has no real world interpretation, only definitional axioms)
December 1, 2024 at 11:13 AM