Decoding Looped Transformers Better for (Almost) Free
cs.LG
Submitted: 2026-10-01
Updated: 2026-10-01
Comments: 32 pages, 19 figures
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Program Synthesis with Large Language Models
- Training-Free Looped Transformers
- Evaluating Large Language Models Trained on Code
- Analog Bits: Generating Discrete Data using Diffusion Models with Self-Conditioning
- Reasoning Beyond Language: A Comprehensive Survey on Latent Chain-of-Thought Reasoning
- Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge
- Training Verifiers to Solve Math Word Problems
- Bridging the Gap Between Latent and Explicit Reasoning with Looped Transformers
- Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach
- Classifier-Free Diffusion Guidance
- Contrastive Decoding Improves Reasoning in Large Language Models
- Stay on topic with Classifier-Free Guidance
- Trusting Your Evidence: Hallucinate Less with Context-aware Decoding
- Thinking by Subtraction: Confidence-Driven Contrastive Decoding for LLM Reasoning
- LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling
- Beyond Parameters: Exploring Virtual Logic Depth for Scaling Laws
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks