Loopholing: pushing self-conditioning all the way in discrete diffusion language models
NandoDF · x · 2026-10-24
NandoDF boosts Adrien Deschenaux's thread on "Loopholing All the Way" for language modelling. Discrete diffusion models sample tokens independently at each step, losing uncertainty from earlier predictions; Loopholing uses self-conditioning to pass a continuous state between steps while still sampling each step. The thread asks whether self-conditioning can be pushed all the way.
Related event: LAW: Language Models Built Entirely on Self-Conditioning(5 posts)→
More from Research
- ICU-REACT paper: teaching LLMs how ICU physicians reason beats pure scaling — SoloGen · 2026-10-24
- Amazon FAR, CMU and Stanford's FIBER trains Unitree G1 to chain skills for delivery runs — CyberRobooo · 2026-10-24
- Dessa's dialog-only voice training isn't new: dGSLM did it in 2022 with just 2k hours — rdesh26 · 2026-10-24
- NYU talk argues 30+ years of consciousness research may be partly wrong — birchlse · 2026-10-24
- Psycholinguistic analysis finds Claude Opus 5 the most out-of-distribution model tested — gleech · 2026-10-24
- Google Research tests 180 agent setups: multi-agent hurts sequential tasks by 39-70% but boosts parallel ones 80.9% — ombre-ai · 2026-10-24