LLM Reasoning as Trajectories: Step-Specific Representation Geometry and Correctness Signals (2026.acl-long)
Copied to clipboard
| Challenge: | Existing models generate tokens by updating high-dimensional representations and decoding from them at each timestep. |
| Approach: | They propose a framework that allows reasoning correction and length control based on derived ideal trajectories. |
| Outcome: | The proposed model can predict correctness and length control based on ideal trajectories. |
Similar Papers
Mapping the Minds of LLMs: A Graph-Based Analysis of Reasoning LLMs (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large Reasoning Models (LRMs) often display unstable behaviors, e.g., hallucinating unsupported premises, overthinking simple tasks, and displaying higher sensitivity to prompt variations. |
| Approach: | They propose a graph-based analytical framework that clusters long, verbose CoT outputs into semantically coherent reasoning steps, then constructs directed reasoning graphs to capture contextual and logical dependencies among these steps. |
| Outcome: | The proposed framework enables quantitative evaluation of internal reasoning structure and quality beyond conventional metrics and provides practical insights for prompt engineering and cognitive analysis of LLMs. |
Explainable Chain-of-Thought Reasoning: An Empirical Analysis on State-Aware Reasoning Dynamics (2025.findings-emnlp)
Copied to clipboard
Sheldon Yu, Yuxin Xiong, Junda Wu, Xintong Li, Tong Yu, Xiang Chen, Ritwik Sinha, Jingbo Shang, Julian McAuley
| Challenge: | Recent advances in chain-of-thought prompting have demonstrated the ability of large language models to perform multi-step reasoning. |
| Approach: | They propose a framework to analyze latent dynamics of CoT trajectories for interpretability . they segment generated CoT into discrete reasoning steps and abstract each step into a spectral embedding based on token-level Gram matrices . |
| Outcome: | The proposed framework segments generated CoT steps into discrete reasoning steps, abstracts each step into a spectral embedding based on token-level Gram matrices, and clusters these embeddements into semantically meaningful latent states. |
Learning Planning-based Reasoning by Trajectories Collection and Process Reward Synthesizing (2024.emnlp-main)
Copied to clipboard
| Challenge: | Recent studies have raised concerns regarding the hallucination and flaws in their reasoning process. |
| Approach: | They propose a framework to learn planning-based reasoning through Direct Preference Optimization on collected trajectories, which are ranked according to synthesized process rewards. |
| Outcome: | The proposed model surpasses GPT-3.5-Turbo on logical reasoning benchmarks on a set of logically-based reasoning tasks. |
Current Advances in LLM Reasoning (2026.acl-tutorials)
Copied to clipboard
| Challenge: | This tutorial examines comprehensive evaluation strategies to assess the reasoning abilities of large language models (LLMs) advanced inference time methods and post-training methods that aim to make LLMs think more like humans are discussed in this tutorial. |
| Approach: | This tutorial explores comprehensive evaluation strategies to assess the reasoning abilities of large language models (LLMs) and discusses two types of methods to improve models’ reasoning: advanced inference time methods, structured and self-improvement inference methods, and post-training methods, such as RLHF, DPO, and GRPO. |
| Outcome: | This tutorial examines evaluation strategies to assess the reasoning abilities of large language models and discusses two types of methods to improve models’ reasoning. |
Beyond Lines and Circles: Unveiling the Geometric Reasoning Gap in Large Language Models (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Recent advances in Large Language Models (LLMs) demonstrate increasing proficiency in complex mathematical and algorithmic tasks, yet their geometric reasoning skills are underexplored. |
| Approach: | They propose a framework that enhances LLMs’ reasoning potential through a multi-agent system conducting internal dialogue. |
| Outcome: | The proposed framework enhances LLMs’ reasoning potential through a multi-agent system conducting internal dialogue. |
Beyond Rejection Sampling: Trajectory Fusion for Scaling Mathematical Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) fine-tuned using rejection sampling retain only correct reasoning trajectories . however, this paradigm treats supervision as a binary filter that systematically excludes teacher-generated errors, leaving a gap in how reasoning failures are modeled during training. |
| Approach: | They propose a fine-tuning strategy that reframes rejection sampling as a structured supervision construction process. |
| Outcome: | The proposed approach outperforms RFT on multiple math benchmarks while retaining only correct reasoning trajectories. |
Discovering a Shared Logical Subspace: Steering LLM Logical Reasoning via Alignment of Natural-Language and Symbolic Views (2026.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to multistep logical reasoning are limited by natural language refinement or external symbolic solvers. |
| Approach: | They propose a logical subspace that captures logical reasoning capabilities in LLMs that are shared across views while remaining independent of surface forms. |
| Outcome: | The proposed approach improves accuracy by 11 percentage points and generalizes well on out-of-domain problems. |
Step Guided Reasoning: Improving Mathematical Reasoning using Guidance Generation and Step Reasoning (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to improve mathematical reasoning require extensive datasets for training or depend on few-shot methods that compromise computational accuracy. |
| Approach: | They propose a training-free adaptation framework that efficiently equips general-purpose pre-trained language models with enhanced mathematical reasoning capabilities. |
| Outcome: | The proposed framework outperforms Qwen2.5-72B-Math-Instruct on MMLU-STEM with a score of 90.9%, compared to 87.3%. |
A Comprehensive Survey of Process Reward Models: Data Generation, Model Construction, and Usage (2026.acl-long)
Copied to clipboard
Congmin Zheng, Jiachen Zhu, Zhuoying Ou, Yuxiang Chen, Kangning Zhang, Rong Shan, Zeyu Zheng, Mengyue Yang, Jianghao Lin, Yong Yu, Weinan Zhang
| Challenge: | Large Language Models (LLMs) have advanced reasoning ability, yet conventional alignment remains dominated by outcome reward models that judge only final answers. |
| Approach: | They summarize applications across math, code, text, multimodal reasoning, robotics, and agents . goal is to clarify design spaces, reveal open challenges, and guide future research toward fine-grained, robust reasoning alignment. |
| Outcome: | The proposed model enables finer credit assignment, richer diagnostics, and improved robustness. |
Truth as a Trajectory: What Internal Representations Reveal About Large Language Model Reasoning (2026.acl-long)
Copied to clipboard
| Challenge: | Existing explainability methods for Large Language Models treat hidden states as static points in activation space, but they are saturated with polysemantic features. |
| Approach: | They propose a framework that shifts analysis from static activations to layer-wise geometric displacement. |
| Outcome: | The proposed framework outperforms existing explainability methods on commonsense reasoning, question answering, and toxicity detection benchmarks. |