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 proposes a graph-constrained approach to scale multi-hop training data by decoupling path discovery from path verbalization, significantly expanding the usable corpus size for LLMs.

The paper introduces MAVEN, a lightweight symbolic reasoning scaffold that significantly improves the generalization and end-to-end success rate of large language models in complex, multi-step tool-ca…

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…

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…

LegalGraphRAG introduces a multi-agent, hierarchical graph retrieval-augmented generation framework to overcome the limitations of traditional RAG in legal domains, achieving state-of-the-art reliable…

The paper proposes GRiD, a novel framework that uses a two-phase training strategy (supervised pre-training and RL fine-tuning) to discover complex, graph-like rules for knowledge graph reasoning, ove…

The paper introduces Contrastive Reflection (CORE), a novel non-parametric method that rapidly improves language model reasoning by distilling contrasts between successful and unsuccessful problem att…

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 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 Geometric Latent Reasoning (GLR), a method that models reasoning as continuous paths in the embedding space, showing that this continuous approach allows LLMs to solve problems us…

The paper introduces StreamMA, a streaming multi-agent reasoning system that significantly reduces latency and improves effectiveness by passing reasoning steps to downstream agents as they are genera…

This paper evaluates the causal reasoning abilities of large language models and finds that they rely heavily on lexical pattern matching rather than structural reasoning.

The paper argues that current embodied planning benchmarks prioritize superficial language prediction over true physical reasoning, introducing new benchmarks and a large-scale dataset to demonstrate…

The paper evaluates LLM reasoning on Boolean satisfiability (SAT) problems, concluding that conventional metrics are misleading and proposing a paired-formula protocol with Accurate Differentiation Ra…

The paper introduces CosmicFish-HRM, a compact language model that achieves adaptive reasoning by dynamically allocating computational effort through a Hierarchical Reasoning Module (HRM), showing tha…

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…

BADGER is a unified, production-grade evaluation framework that integrates text-to-SQL assessment with agentic behavior evaluation, significantly outperforming existing benchmarks on industry queries.