Search: “reasoning”

Reducing Cognitive Overhead in Tool Use via Multi-Small-Agent Reinforcement Learning

arXiv:2508.08882v5 Announce Type: replace Abstract: Recent advances in multi-agent systems highlight the potential of specialized small agents that collaborate via division of labor. Existing tool-int

agentsarxiv-cs-ai

2 Oct 2026

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Research

Ten years ago, AlphaGo’s Move 37 shocked the world. It wasn’t intuition alone that produced it. AlphaGo could search possible futures, test …

DGX agent

Ten years ago, AlphaGo’s Move 37 shocked the world. It wasn’t intuition alone that produced it. AlphaGo could search possible futures, test its instincts and reason about what would happen next. In a

researchyann-lecun--x

2 Oct 2026

Tutorials

UniWAM: Unified World-Action Model

arXiv:2610.02054v1 Announce Type: new Abstract: Vision-language-action models benefit from the understanding and reasoning capabilities of pretrained vision-language models, but action-only supervisio

tutorialsarxiv-cs-ro

2 Oct 2026

Video Evidence Indexing: Learning Where to Look from Video Previews for Token-Budgeted Long-Video Question Answering

arXiv:2610.00757v1 Announce Type: new Abstract: Long-video question answering is limited by the high cost of visual tokens and by the fixed context width of current VLMs. A long-video question may req

local-aiarxiv-cs-cv

2 Oct 2026

A Tilted Bowl Is Not a Slippery Slope: Compressing Looped Models

arXiv:2609.39277v1 Announce Type: cross Abstract: Looped models reason by applying the same block of weights many times, so compressing that block saves memory traffic on every loop. Compressed looped

researcharxiv-cs-cl

Better Supervision Is Nearby: Neighborhood On-Policy Self-Distillation

model-releasesarxiv-cs-cl

arXiv:2609.39687v1 Announce Type: new Abstract: On-policy self-distillation (OPSD) trains mathematical reasoning models using a privileged teacher that sees a reference solution and supervises student

Debias It Yourself: Teaching LLMs Cognitive Bias Mitigation Interventions

arXiv:2609.40124v1 Announce Type: new Abstract: Bias has long been studied in social psychology and cognitive science, where decades of research have produced a body of validated interventions that re

safetyarxiv-cs-cl

DynaHarness: A Dynamic Physical Harness for Self-Evolving Robot Agents

arXiv:2609.40306v1 Announce Type: cross Abstract: Pretrained robot policies provide useful action priors, but long-horizon manipulation still requires coordination between semantic reasoning and physi

Explicit Trajectory Diversity for RL-Based Post-Training of LLM Agents

arXiv:2609.38805v1 Announce Type: cross Abstract: LLM agents often admit multiple high-quality solutions to the same task, differing in reasoning structure, tool-use pattern, or interaction trajectory

From Charts to Code: A Hierarchical Benchmark for Multimodal Models

arXiv:2510.17932v5 Announce Type: replace-cross Abstract: We introduce Chart2Code, a new benchmark for evaluating the chart understanding and code generation capabilities of large multimodal models (L

GRPO Training Dynamics for Small Language Models

arXiv:2609.39321v1 Announce Type: cross Abstract: Group Relative Policy Optimization (GRPO) has emerged as a memory-efficient reinforcement fine-tuning (RFT) technique for reasoning-intensive tasks. H

KilometerVision: A New Frontier for Large-Scale Spatial Intelligence in VLMs

model-releasesarxiv-cs-cv

arXiv:2609.39588v1 Announce Type: new Abstract: We push the frontier of large-scale spatial intelligence in Vision-Language Models (VLMs) and introduce the first benchmark that probes geographical lay

Less Data Approximates More: Earning Faithful Confidence in High-Stakes Domains

model-releasesarxiv-cs-lg

arXiv:2604.08454v2 Announce Type: replace Abstract: Large language models are increasingly deployed in high-stakes domains, where confident yet incorrect inferences may cause severe real-world harm, b

MADBench: Benchmarking the Security of Multi-Agent Debate

arXiv:2609.39146v1 Announce Type: new Abstract: Multi-agent debate (MAD) can improve large language model (LLM) reasoning by allowing multiple agents to exchange and critique their answers to the same

Multi-LLM Collaborative Alignment via Stackelberg Games

arXiv:2609.39076v1 Announce Type: new Abstract: A pool of language models can collaborate and improve collectively by learning from one another's responses. These interactions depend on the instructio

Taming Speculative Search for Test-Time Scaling in LLM Serving

arXiv:2609.39334v1 Announce Type: cross Abstract: Test-time scaling has recently emerged as a powerful approach for improving LLM reasoning by allocating additional computation during inference, subst

researcharxiv-cs-cl

Unlearnable, or Unmeasured? On the Reliability of Difficulty Labels in RLVR

arXiv:2609.40115v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) has become an important approach for improving reasoning during post-training. Recent work suggest

researcharxiv-cs-ai

When Order Matters: First-Speaker Bias and Mitigation through Personality in Sequential Multi-Agent Debate

arXiv:2609.38964v1 Announce Type: new Abstract: Multi-agent debate (MAD) is often used to improve large language model (LLM) reasoning, but sequential debate is rarely a neutral aggregator of agents'

A Skill-augmented Agentic Framework and Benchmark for Multi-Video Understanding

model-releasesarxiv-cs-cv

arXiv:2603.14733v2 Announce Type: replace Abstract: Multimodal Large Language Models have achieved strong performance in single-video understanding, yet their ability to reason across multiple videos

Tutorials

Abductive World Modeling via Causal Representation Learning

arXiv:2609.36985v1 Announce Type: cross Abstract: The central challenge of world modeling is to learn representations that capture how the world evolves. However, existing world models predominantly r

tutorialsarxiv-cs-ai

From Dissonance to Orchestration: Teacher Intervention in On-Policy Distillation

model-releasesarxiv-cs-cl

arXiv:2609.37510v1 Announce Type: new Abstract: On-policy distillation (OPD) trains a student on its own reasoning trajectories using feedback from a stronger teacher. Teacher interventions can improv

Layer-Informed Fine-Tuning via Three-Stage Functional Segmentation of LLMs

arXiv:2609.38027v1 Announce Type: new Abstract: In recent years, the performance of large language models (LLMs) on reasoning tasks has been remarkable, even surpassing human capabilities on various b

researcharxiv-cs-cl

Learning from Think-Mode Advantage via On-Policy Distillation

arXiv:2609.37044v1 Announce Type: new Abstract: Explicit intermediate reasoning gives large language models (LLMs) a stronger problem-solving mode. We study learning from this think-mode advantage via

safetyarxiv-cs-cl

Privy to the Foil: Recasting Value Estimation with a Self-Privileged Critic for RLVR

arXiv:2609.37825v1 Announce Type: cross Abstract: Assigning credit to intermediate steps remains a central challenge in training Large Language Models (LLMs) on multi-step reasoning tasks with sparse

RESCUE: Repairing Language Model Errors to Sparse Circuits via Reinforcement Learning

arXiv:2609.36813v1 Announce Type: new Abstract: Large language models (LLMs) exhibit strong general capabilities that mechanistic interpretability has attributed to sparse computational circuits. Howe

local-aiarxiv-cs-lg

Dr. MAS: Stable Reinforcement Learning for Multi-Agent LLM Systems

arXiv:2602.08847v2 Announce Type: replace Abstract: Multi-agent LLM systems enable advanced reasoning and tool use via role specialization, yet reliable reinforcement learning (RL) post-training for s

agentsarxiv-cs-lg

IQuestLab/IQuest-Q1 · Hugging Face

IQuest-Q1 is a Mixture-of-Experts (MoE) model developed by IQuest for agentic coding, reasoning, and multi-step tool use. It comprises approximately 320B total parameters, with an estimated 15B parame

local-air-localllama

M3OS: A Monte Carlo Graph Search-Orchestrated Multi-Agent LLM System for Evidence-Traced Molecular Optimization

arXiv:2609.34491v1 Announce Type: new Abstract: Small-molecule optimization integrates medicinal-chemistry reasoning and computational evidence through iterative, multi-objective decisions. When large

agentsarxiv-cs-lg

On-Policy Attention Linearization

arXiv:2609.31947v1 Announce Type: new Abstract: Hybrid transformer architectures that replace most softmax attention layers with linear attention offer transformer-level quality at a fraction of the m

safetyarxiv-cs-lg

On Temporal Binding in Large Audio Language Models

arXiv:2609.34806v1 Announce Type: cross Abstract: Reasoning about temporal structure of audio recordings requires Large Audio Language Models (LALMs) to associate sound events with their temporal posi

researcharxiv-cs-lg

RAM Offloading with vLLM - tcclaviger appreciation post

model-releasesr-localllama

Thanks to tcclaviger, vLLM now has expert RAM offloading support (link). This makes frontier models much more accessible on a local setup! I was able to run the original DeepSeek-V4-Flash-Vision-Exp o

ReSight-SMC: Two-Stage Power Sampling via Island SMC with Visual Scouts

model-releasesarxiv-cs-cv

arXiv:2609.34905v1 Announce Type: new Abstract: Power sampling has emerged as a training-free approach to LLM reasoning, eliciting capabilities comparable to reinforcement learning by sharpening the m

SCOPD: Sparse-Context On-Policy Self-Distillation for Efficient Vision-Language Models

arXiv:2609.34044v1 Announce Type: new Abstract: Reasoning vision-language models (VLMs) process images and videos as long sequences of visual tokens, making inference expensive. Training-free token pr

safetyarxiv-cs-cv

Selecting Diverse SFT Traces Improves Post-RL Generalization

arXiv:2609.33780v1 Announce Type: new Abstract: Verified solutions are not equally useful for preparing reasoning models for reinforcement learning (RL). We present a comprehensive study of route dive

researcharxiv-cs-lg

Do LLMs Understand Context? A Knowledge Graph-Based Evaluation Framework

arXiv:2609.30484v1 Announce Type: new Abstract: While large language models (LLMs) have achieved remarkable linguistic capabilities, a profound question lingers at their core: do these models truly co

researcharxiv-cs-ai

Evolutionary Safety of Recursive Self-Improving AI: Taxonomy, Risk Discovery, and Evaluation

arXiv:2609.31186v1 Announce Type: new Abstract: Artificial intelligence is advancing rapidly, with increasingly capable systems taking larger roles in reasoning, decision-making, scientific discovery,

Governed Deduction: Policy-Grounded Premise Authorization Beyond Relevance

arXiv:2609.31029v1 Announce Type: new Abstract: Reasoning systems usually treat premise use as a question of relevance: if a fact is available and useful, it may be selected for inference. Authorizati

Highlight-Then-Summarize: Learning to Compress Evidence for Long-Context Understanding

arXiv:2609.31382v1 Announce Type: cross Abstract: Long-context understanding requires large language models (LLMs) to reason over lengthy documents, conversations, and code, yet task-relevant evidence

When Is a Multi-Agent Code Judge Actually Grounded? Two Label-Free Measurements, and a Judge That Declines to Guess

arXiv:2609.30328v1 Announce Type: new Abstract: When one language model judges whether another's code is correct, it does not report the absence of evidence. It returns a confident verdict with reason

agentsarxiv-cs-ai

2026 in LLMs (so far)

model-releasessimon-willison

On Friday I gave the closing keynote at the WeAreDevelopers World Congress North America in San Jose. I tied together the key trends from the past year into a chronological exploration of everything t

27 Sept 2026

Improved and fixed template for GPT-OSS (again). Includes preserve_thinking and fix for Unsloth-induced bug

I posted an updated GPT-OSS template a couple of months ago, which was based on Unsloth's version. It turns out that both Unsloth's version (and, thus, mine) contain a very serious bug that can degrad

local-air-localllama

26 Sept 2026

GPT-6-Astra Lights Up Embodied Navigation: Evaluation in Zero-Shot Vision-and-Language Navigation in Continuous Environments

model-releasesarxiv-cs-ro

arXiv:2609.29861v1 Announce Type: new Abstract: We investigate whether GPT-6-Astra, a general-purpose foundation model, can navigate unfamiliar environments using its own perception, reasoning, and de

25 Sept 2026

HarnessPAI: An Evolving Harness for Physical AI

arXiv:2609.29166v1 Announce Type: cross Abstract: Physical AI aims to build embodied agents that perceive the world, understand and reason about it, and decide how to act. Yet the field has focused pr

25 Sept 2026

Qwen3.8-27B: Using KV Cache Transplants to Boost Output Quality

model-releasesr-localllama

Since my last post, I've been thinking about different options for dynamic performance degradation, trying to squeeze as much high-quality inference out of my GPU as I can. Over the weekend I read thi

25 Sept 2026

DCRL: Decoupling and Coupling Reinforcement Learning via Policy-Reward Manifold Alignment

arXiv:2609.27572v1 Announce Type: cross Abstract: Reinforcement learning (RL) has emerged as a key paradigm for improving the reasoning capabilities of large language models (LLMs). However, existing

Do We Need Complex Topology Control? Distinct-Peer Random Routing Improves Cost-Efficiency in Sparse Multi-Agent Debate

arXiv:2609.27150v1 Announce Type: new Abstract: Multi-agent debate (MAD) has emerged as a promising paradigm for improving the reasoning accuracy of large language models (LLMs) through iterative peer

Energy-guided Recursive Model

arXiv:2607.10128v3 Announce Type: replace Abstract: Recursive models show promise on reasoning and language tasks, yet their test-time scaling lacks a principled criterion for selecting trajectories o

local-aiarxiv-cs-lg

UkisAI Swift Series / 27B, Flash Next and Bonsai 2 + GSQ-RCO / -63.4% thinking, x1.95 speed with xhigh accuracy

model-releasesr-localllama

Hey everyone, Jovan from UkisAI here! Today, we are introducing Swift, a family of efficient reasoning LLMs based on Qwen, trained by penalizing tokens related to pathological overthinking patterns an