This work introduces Structured Thoughts, a framework that organizes reasoning into alternating blocks and blocks that captures exploratory scratch work, while the distilled conclusion of that step contains the distilled conclusion of that step.
Abstract
Large language models (LLMs) excel at generating long chains of thought, but long reasoning traces are often verbose and memory-inefficient. In this work, we introduce Structured Thoughts, a framework that organizes reasoning into alternatingandblocks:captures exploratory scratch work, whilecontains the distilled conclusion of that step. We construct a dataset of structured thoughts by segmenting reasoning traces intoblocks and prompting an LLM to summarize each step into its corresponding. Fine-tuning pretrained foundation models on this reformatted data produces models that adopt the structured reasoning style, leading to performance gains of up to 8.08\% on reasoning benchmarks compared to standard SFT. The explicit structure also enables context pruning: after each/pair, thecan be pruned, allowing the model to retain conclusions without keeping the full scratch work in the context. A proof-of-concept pruning implementation achieves an average of 85\% memory / context savings with an 8.67\% performance drop across mathematical tasks.
Memory-Augmented Compression is proposed, a training-free framework that constructs reusable reasoning memories from historical traces and retrieves them as prefill-side scaffolds to compensate for information lost during compression.
Si-Meng Zhang, Yilong Chen, Wenyuan Zhang et al.· 0 citations
This work proposes ChainPrune, a novel reasoning path semantic structural optimization method to efficiently and controllably synthesize self-generated high-quality training data and incorporates a DPO-based preference learning method combined with supervised loss, effectively mitigating false reward suppression.
Weihang Pan, Zhengxu Yu, Yuxiang Zhang et al.· 1 citation
Reason Popper-ly, a neurosymbolic framework that uses inductive logic programming (ILP) to learn relation composition rules from reasoning traces and deploys them as an online verifier for step-level correction, consistently improves terminal accuracy over standard CoT.
A novel fragment-based reasoning framework is introduced in which the model first extracts parallel source-target fragments from retrieved similar exemplars, and uses these fragments as intermediate reasoning traces to produce the final translation.
Maxime Bouthors, J. Crego, François Yvon· 0 citations
DynaRule is proposed, an end-to-end framework that injects the given rules into the KV cache and turns retrieval into an internal, learnable, step-wise process, and can re-attend to the most relevant rules at each step, dynamically replacing outdated ones to support more stable multi-step reasoning.
Bohan Yu, Pengfei Cao, Chen Han et al.· 1 citation
Large language models (LLMs) are often post-trained on pre-collected reasoning trajectories to improve their reasoning capability. Such trajectories tend to be long due to complex, interwoven paths, which often include detours on the path toward the answer. However, it has been underexplored whether LLMs indeed benefit from learning complete trajectories in post-training, such as supervised fine-tuning (SFT). Starting from our pilot study, we find that full trajectories provide only limited benefit, while partial trajectories are effective even under heavy truncation. We analyze redundancy in reasoning trajectories through attention-based analyses and controlled token-removal studies, both of which show that intermediate tokens contribute minimally to final reasoning quality. This suggests that avoiding redundant information may allow LLMs to internally infer coherent alternatives by inferring missing steps from their internal knowledge, given known trajectory endpoints. Furthermore, we show that training LLMs using endpoints leads to consistent changes in reasoning behavior, and that it also benefits post-training methods based on reinforcement learning or on-policy distillation, highlighting the need to revisit complete reasoning traces. Code is available at https://github.com/naver-ai/revisiting-trace.
Jaehui Hwang, Sangdoo Yun, Byeongho Heo et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
