frontier models
Earlier Loops Made the Final Loop Better
LoopCD contrasts the final prediction of a looped transformer with an earlier recurrent pass without training another model.
Summary
LoopCD contrasts the final prediction of a looped transformer with an earlier recurrent pass without training another model.
Looped transformers expose aligned weak and strong predictions because the same block produces decodable states at successive passes. LoopCD uses an earlier pass as a contrastive signal for the final prediction. The authors report AIME 2024 pass@1 rising from 61.88 to 73.33 percent for Ouro-2.6B-Thinking and HumanEval pass@1 from 22.56 to 31.71 percent for Huginn, while some configurations matched full-depth baselines with 22.5 to 48.2 percent fewer forward FLOPs.
Why it matters
LoopCD contrasts the final prediction of a looped transformer with an earlier recurrent pass without training another model.
Limits and context
No additional limitation was separately recorded.
Key claims
LoopCD contrasts the final prediction of a looped transformer with an earlier recurrent pass without training another model.
Evidence: source-2026-10-03-011
Sources
- arXiv preprint 2610.02185arXiv · primary research
Corrections
No corrections have been recorded for this story.