Paper

Self-Regulating Annealing in Heavy-Tailed Diffusion Models

Unreadunread

AI summary

arXiv:2606. 01645v1 Announce Type: new Abstract: Diffusion models have emerged as a leading framework for deep generative modeling.

Reading status

Priority

About this paper

Authors: Keito Wakatsuki, Hideaki Shimazaki
Released: 2026-06-02 (2mo ago)
Topic: Related to current project
Relation to my work: Read next because Self-Regulating Annealing in Heavy-Tailed Diffusion Models overlaps with clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "The marker is a representational handle, not a behavioural one — sharing it between a villain persona and the assistant transfers no misalignment (HIGH confidence)", experiment "Follow-up to #354: cascading chunk-binding — does A→B, B→C, C→D propagate the full chain on a recipient trained only to emit A?". Matching terms: latin, rate, full, model. Source: arxiv stat.ML (Machine Learning).
arXiv: 2606.01645

Ask Claude about this paper

Posts a comment with @claude so the runner spins up a Claude Code session scoped to Self-Regulating Annealing in Heavy-Tailed Diffusion Models. Replies appear in the conversation below.

Comments

No comments yet.

Related records

No links yet.