Paper

When Does LeJEPA Learn a World Model?

Unreadunread

AI summary

arXiv:2605. 26379v1 Announce Type: new Abstract: A representation that scrambles the true degrees of freedom of the world cannot support reliable planning or compositional generalization.

Reading status

Priority

About this paper

Authors: David Klindt, Yann LeCun, Randall Balestriero
Released: 2026-05-27 (1mo ago)
Topic: Related to current project
Relation to my work: Read next because When Does LeJEPA Learn a World Model? 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 "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Coupling evil personas with wrong answers fails to protect Qwen2.5-7B from EM-induced alignment collapse — and the apparent capability ordering across coupling conditions is mostly eval contamination (LOW confidence)". Matching terms: class, rect, under, alignment, distributional, line, recipe, control. Source: arxiv stat.ML (Machine Learning).
arXiv: 2605.26379

Ask Claude about this paper

Posts a comment with @claude so the runner spins up a Claude Code session scoped to When Does LeJEPA Learn a World Model?. Replies appear in the conversation below.

Comments

No comments yet.

Related records

No links yet.