titleDecoding Looped Transformers Better for (Almost) Free
authorsWeihao Liu, Huangjie Zheng, Tianrong Chen, Rohit Dilip, Richard He Bai, Yizhu Jiao, Yuyang Wang, Ruixiang Zhang
date2026-10-01
venuearXiv (preprint, not peer reviewed)
idarXiv:2610.02185
affiliationApple
upvotes23 (Hugging Face Daily Papers, 2026-11-01)
linksabs pdf hf
abstractLooped Transformers achieve parameter efficiency by repeatedly executing a shared block across recurrent loops. Each loop yields an intermediate representation decodable for the same next token, yet standard decoding discards earlier states. Because earlier loops embody less computation, recurrence inherently supplies aligned weak-and-strong prediction pairs without auxiliary models or external training. We introduce LoopCD, a training-free contrastive decoding framework that guides token selection by contrasting the final prediction with an earlier recurrent pass, operating either in logit space with one extra output pass (LoopCD-Logits) or in hidden-state space with zero output overhead (LoopCD-Hidden). Across four looped Transformer families, LoopCD delivers substantial, consistent gains at full recurrent depth: LoopCD-Logits raises Ouro-2.6B-Thinking's AIME 2024 pass@1 from 61.88% to 73.33%, while LoopCD-Hidden lifts Huginn's HumanEval pass@1 from 22.56% to 31.71%. Crucially, these performance gains enable halving the number of recurrent loops while still matching or exceeding full-depth unguided baselines, reducing forward FLOPs by 22.5% to 48.2%. By transforming intermediate recurrent states into effective guidance signals, LoopCD achieves superior decoding quality while substantially reducing inference compute.

cite

@misc{liu2026decoding,
  title={Decoding Looped Transformers Better for (Almost) Free},
  author={Weihao Liu and Huangjie Zheng and Tianrong Chen and Rohit Dilip and Richard He Bai and Yizhu Jiao and Yuyang Wang and Ruixiang Zhang},
  year={2026},
  eprint={2610.02185},
  archivePrefix={arXiv},
  url={https://arxiv.org/abs/2610.02185}
}