The paper proposes a hybrid reasoning framework where Large Language Models (LLMs) generate code to encode complex optimization problems into a preference-based Maximum Satisfiability (MaxSAT) format,…

The paper introduces HRBench, a unified and comprehensive evaluation framework for systematically benchmarking and comparing various thinking-mode switching strategies in hybrid-reasoning LLMs.

The paper introduces LinTree, a method that explicitly structures the search history of LLM reasoning traces using parent pointers, significantly improving task performance and search efficiency compa…

The paper introduces an Integrated, cross-Architecture Reasoning (IAR) framework to provide a unified and robust method for interpreting the opaque reasoning processes within Large Language Models.

The paper proposes a hybrid SAT-solving framework that uses a probabilistic-bit (p-bit) Ising sampler to guide Conflict-Driven Clause-Learning (CDCL) solvers, significantly reducing internal search ef…

PlanarBench introduces a novel benchmark to test LLM spatial reasoning by requiring them to draw planar graphs as ASCII art from an edge list, finding that edge count is a stronger difficulty predicto…

The paper analyzes the failure modes of aggressive 2-bit quantization in large reasoning models, proposing lightweight controls like FP16 planning and loop rescue to restore accuracy and achieve pract…

This paper localizes the attention heads within LLMs responsible for specific reasoning steps, finding that specialized heads handle factual retrieval while higher layers manage global information int…

The paper introduces an automatic numeric-remapping attack to test the robustness of LLMs on arithmetic word problems, finding that LLMs remain sensitive to small numeric changes in datasets like GSM8…

DenseSteer is a training-free inference-time framework that improves the math reasoning capabilities of small language models by steering their internal representations toward a 'Dense Reasoning' patt…

The paper introduces TRAILS~, a novel method that improves code correctness validation by grounding LLM reasoning in concrete (input, output) pairs derived from specifications, achieving state-of-the-…

The paper introduces Reasoning in Memory (RiM), a latent reasoning method that replaces autoregressive token generation with fixed memory blocks to enable compute-efficient internal working memory for…

The paper proposes a trust-boundary architecture using Lean 4 to verify the deterministic structured computations surrounding LLM pipelines, providing verifiable certificates for high-stakes deploymen…

This paper introduces the Data-Model Compatibility (DMC) metric to quantify how suitable a dataset is for reasoning distillation, showing that optimizing data selection using DMC significantly improve…

The paper introduces Chunk-Level Guided Generation, a training-free method that uses an off-the-shelf large language model (LLM) as a process scorer to guide small model generation, achieving performa…

The paper introduces BenGER, a comprehensive benchmark for evaluating LLMs on German legal reasoning, demonstrating that closed-flagship models perform best and that human-AI co-creation significantly…

The paper introduces DPrivBench, a new benchmark to test whether large language models (LLMs) can automate the complex reasoning required to verify differential privacy guarantees for algorithms.

This paper investigates how different types of compressed reasoning data (Explicit, Composed, Implicit CoT) affect LLM performance during post-training, finding that the choice of compression and subs…

The paper introduces a novel framework to quantify faithful confidence expression (FC) in Large Reasoning Models (LRMs), finding that FC remains a significant and challenging reliability target for th…