Search: “reasoning”

arXiv:2610.11215v1 Announce Type: new Abstract: Reconstructing complete, scene-aligned 3D objects from casual images requires integrating sparse, uncertain observations and inferring surfaces hidden b

hardwarearxiv-cs-cv

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

GroundSight at GroundLM 2026 Shared Tasks: GoldenViewVQA

DGX agent

arXiv:2610.11402v1 Announce Type: new Abstract: GoldenViewVQA requires models to jointly answer driving-scene questions and identify the camera view containing the supporting visual evidence, making p

model-releasesarxiv-cs-cv

9 Oct 2026

How Should a Prompt Optimizer Spend a Tight Budget? BudgetAPO with Noise-Adaptive Evaluation

arXiv:2610.05671v2 Announce Type: replace Abstract: Automatic prompt optimization (APO) has been widely employed to adapt large language models without updating their weights, yielding promising resul

researcharxiv-cs-ai

iAm.md: Robot Skill Self-Assessment through Agentic Introspection for Unknown Open-Vocabulary Domains

arXiv:2610.10962v1 Announce Type: cross Abstract: Agentic AI based on Large Language Model generalization capabilities offers a wide range of potential applications, including planning for embodied ta

agentsarxiv-cs-ai

Just another purely open-weight models benchmark

Live results, coding benchmarks included, agentic benchmarks included, domains and other classifications filters included. (Spoiler: DeepSeek-V4-Vision-Exp rules, but other models have their rule area

KDFP: A first-principles approach to knowledge distillation in large language models

model-releasesarxiv-cs-lg

arXiv:2610.10854v1 Announce Type: new Abstract: Knowledge distillation is an established technique for improving the capabilities of small, efficient student models by training them with the represent

Language Modeling is Monotone Compression

arXiv:2610.11031v1 Announce Type: cross Abstract: A long-standing hypothesis in artificial intelligence and neuroscience posits that intelligence is closely related to compression: the ability to comp

researcharxiv-cs-ai

Hardware

Language Models as AI Research World Models

arXiv:2610.12235v1 Announce Type: new Abstract: AI research agents automate the cycle of proposing, implementing, and evaluating experiments, opening a path toward recursive self-improvement. Yet thei

hardwarearxiv-cs-cl

MetaEncoder: Exploring the Limit of Bi-Encoders for Multimodal System One Decision Making with Natural Language Interface

arXiv:2610.11316v1 Announce Type: cross Abstract: System One models output constrained decisions and probability distributions rather than free-form text generation. While prevailing paradigms rely on

MetaOPD: Meta-Learned Token Weighting for On-Policy Distillation

arXiv:2610.11989v1 Announce Type: new Abstract: On-policy distillation (OPD) trains a student on its own generated responses using token-level teacher supervision. However, uniform weighting overlooks

MotiveMob: Motivation as Semantic Action for Closed-Loop Human Mobility Generation

arXiv:2610.11442v1 Announce Type: new Abstract: Human mobility generation, an important task in urban research, synthesizes trajectory data for urban planning and transportation management. Human mobi

researcharxiv-cs-lg

MSGAT: Multi-Head Spiking Graph Attention with Similarity-Space Fusion for Image-Text Retrieval

arXiv:2610.11526v1 Announce Type: cross Abstract: Spiking neural networks (SNNs) offer an energy-efficient computing paradigm through sparse event-driven computation, showing great potential for effic

[Paper] EngramEdit: Decoupled Knowledge Updates in LLMs through Conditional Memory

Conditional memory architectures such as DeepSeek Engram use input n-grams to look up learned embeddings, expanding the capacity of large language models (LLMs) with limited additional computation. Be

PearlVLA: Progressive Embodied Action-Plan Refinement in Latent Space

arXiv:2606.17924v2 Announce Type: replace-cross Abstract: Current Vision-Language-Action (VLA) models face a trade-off between efficient action generation and explicit deliberation. Directly decoding

PlanWAM: Planning-Shaped Future Representations for End-to-End Autonomous Driving

arXiv:2610.11382v1 Announce Type: cross Abstract: World models in end-to-end autonomous driving predict future scene evolution to provide foresight for trajectory planning. Existing methods mainly stu

agentsarxiv-cs-lg

Pose-Free Feed-Forward 3D Inpainting via Learnable Mask Attention and Support Token Refinement

arXiv:2610.11857v1 Announce Type: new Abstract: 3D scene inpainting aims to recover missing or occluded regions in edited 3D scenes, while ensuring geometric and textural consistency. Existing approac

researcharxiv-cs-cv

Probabilistic Sensing, Deterministic Authority: Admitting Model-Produced Observations into Sufficiency-Checked Governance Contracts

arXiv:2610.10978v1 Announce Type: cross Abstract: When a field that an authority contract needs exists only in unstructured evidence, a model can sense it. We admit the model's output only as an obser

Qwen 3.8 Flash Next-GSQ-RCO-IQ2_XS at ~21 tok/s on just an RTX 3060 12GB + 16GB DDR4 RAM(No gate pruning, 100% bit-exact)

(Don't judge by the screenshot, the cache is cold. It hits 24+ tok/s with a warm cache!) About two months ago, I made a post here asking whether predicting which MoE experts would be used on the next

READ-Bench: Benchmarking Historical Instance Retrieval for Time-Series Diagnosis

arXiv:2609.32123v2 Announce Type: replace Abstract: Time-series diagnostic systems rarely rely on retrieving relevant historical cases, and when they do, retrieval is evaluated only indirectly through

Refine Connections, Close the Gap: A Reliable Enhancement Framework for Driving Scene Topology

arXiv:2610.11058v1 Announce Type: new Abstract: In autonomous driving, understanding scene topology - the connectivity between lanes and traffic elements - is critical for safe path planning and motio

agentsarxiv-cs-cv

SAIL: Scientific Agentic Intelligence via a Science-Aware Loop

arXiv:2610.11451v1 Announce Type: new Abstract: We introduce SAIL, an open model with 35B total and 3B active parameters for literature research, scientific coding, and multi-step research workflows.

safetyarxiv-cs-cl

Same Outcome, Different Evidence: Intent Recovery in LLM Safety Evaluation

arXiv:2610.11766v1 Announce Type: cross Abstract: Safety evaluations of large language models commonly summarize harmful-output behavior with attack success rate (ASR). Yet the same non-harmful outcom

Applications

SPAN-Nav: Generalized Spatial Awareness for Versatile Embodied Navigation

arXiv:2603.09163v2 Announce Type: replace Abstract: Recent embodied navigation approaches leveraging Vision-Language Models (VLMs) demonstrate strong generalization in versatile Vision-Language Naviga

applicationsarxiv-cs-ro

SpikeSSL: A Universal Spike Inference Framework with Dynamics-Informed State-Space Layers

arXiv:2610.11456v1 Announce Type: new Abstract: Two-photon calcium imaging is a standard tool for recording large neural populations in vivo, yet inferring spikes accurately across the growing diversi

Stepping away from Benchmarks and Code, what models are you using for Creating writing projects

All anyone ever talks in this sub is about x new model having x benchmark scores and how well it can code. Or some new tech to increase t/s for qwen. So I wanted to go back to the roots of this sub an

StoreBench: A Live-Commerce Environment for Evaluating and Training Autonomous Operator Agents

arXiv:2610.10942v1 Announce Type: new Abstract: Reinforcement learning environments are now a primary lever for improving large language model (LLM) capabilities in post-training, yet most agentic ben

Strix halo + CMP 170HX setup

A bit of backstory : i got lucky in August and was able to grab a CMP 170 HX 8GB for around 500 on alibaba. I tried later to get a second one for around 1k but my order got cancelled and prices went t

SWE-Journey: Towards More Realistic Evaluation of Coding Assistants through Long-Horizon, Multi-Turn Interaction

arXiv:2610.11559v1 Announce Type: cross Abstract: Coding assistants such as Claude Code and Codex have become a major application of LLM agents, yet existing benchmarks remain far from real-world use,

Tested Mellum2.1-12B-A2.5B on PI Coding Agent - surprisingly usable, but not great at one-shot projects

I tested Mellum2.1-12B-A2.5B (Q8) locally using Pi and llama-server. All five tests were one-shot. Results were pretty mixed: - Pelican SVG, Browser OS, Minecraft: Poor results. - Bouncing Hexagon: Ph

Hardware

The Operator Mismatch Problem: Deploying BEV Perception with Portable GPU Compute

arXiv:2610.11504v1 Announce Type: new Abstract: Modern autonomous driving systems rely on bird's-eye-view (BEV) perception models that fuse camera and LiDAR inputs to detect objects in 3D space. These

hardwarearxiv-cs-ai

TRL v1.15 is out, and it’s an absolute banger of a release for memory-efficient post-training. The main change: SFT, DPO, KTO, GRPO, RLOO an…

model-releasesclem-delangue--x

TRL v1.15 is out, and it’s an absolute banger of a release for memory-efficient post-training. The main change: SFT, DPO, KTO, GRPO, RLOO and Distillation now use a fused LM head by default. Instead o

ttok 1.0

model-releasessimon-willison

Release: ttok 1.0 I released ttok 0.4, ran uv tool upgrade ttok, piped a file into the new version... and realized that it was defaulting to the GPT-4 tokenizer when it should very clearly default to

Universal Textual Teaching for LLMs

arXiv:2610.12114v1 Announce Type: new Abstract: Knowledge distillation (KD) transfers knowledge from stronger Teacher models to weaker Student models, but most methods require training the Student par

Unlocking the Regulatory Genome by ARGUS: An Evidence-Constrained Agentic Framework for Interpreting Single Nucleotide Variants

arXiv:2610.12281v1 Announce Type: cross Abstract: Over 90% of disease-associated variants from genome-wide association studies fall in noncoding regulatory regions, yet their functional interpretation

agentsarxiv-cs-ai

When KL Regularization Misfires in Group Policy Optimization

arXiv:2610.12161v1 Announce Type: cross Abstract: Why does removing reference-policy KL regularization sometimes improve group policy optimization? This motivates studying how reference-policy informa

safetyarxiv-cs-cl

Where Draft Trees Lose Target Mass: Exit-Guided Speculative Decoding

arXiv:2610.11750v1 Announce Type: new Abstract: Tree-based speculative decoding verifies multiple draft continuations in one target-model pass, but finite trees built from draft scores face a fundamen

researcharxiv-cs-ai

Tutorials

Whose Ground Truth? Embracing Ambiguity in Human-Centered AI

arXiv:2610.10805v1 Announce Type: new Abstract: As AI systems increasingly interact with people and make decisions about them, understanding human interpretations becomes an important part of developi

tutorialsarxiv-cs-ai

Why On-Policy Distillation Sometimes Fails: Vanishing Learning Signals

arXiv:2610.11247v1 Announce Type: cross Abstract: On-policy distillation (OPD) enables effective capability transfer between language models, yet the mechanisms underlying its failures are not fully u

A Good Self-Teacher Meets the Student Where They Are: Joint On-Policy Learning and Teaching

arXiv:2610.10447v1 Announce Type: new Abstract: Reinforcement Learning (RL) from outcome rewards suffers from sparse supervision, particularly on difficult, long-horizon tasks where successful traject

safetyarxiv-cs-lg

Agentic Scene Policies

arXiv:2509.19571v2 Announce Type: replace-cross Abstract: Designing or learning robot policies that generalize zero-shot across a range of language instructions and objects is a core problem in roboti

safetyarxiv-cs-cv

Alice: A Large-Scale German Benchmark for Rubric-Based Multi-Dimensional Automatic Short Answer Scoring

arXiv:2610.09661v1 Announce Type: new Abstract: Automatic Short Answer Scoring (ASAS) is central to NLP for Education. However, openly available benchmarks remain scarce, and existing datasets largely

Arctic Questions, Missing Answers: A Dataset and Benchmark for LLM Abstention in Arctic Science

arXiv:2610.09446v1 Announce Type: new Abstract: Large language models (LLMs) should abstain from scientific multiple-choice questions when no option is valid, but frequent abstention alone does not de

Ask the Expert: LLM-Guided Reinforcement Learning for Autonomous Cyber Defense

arXiv:2610.09337v1 Announce Type: cross Abstract: Policy-based reinforcement learning (RL) approaches have produced promising results for autonomous cyber defense; however, they are sample-inefficient

safetyarxiv-cs-lg

BEACON-SP: Ontology-Grounded GraphRAG Framework for Clinical Suicide Risk Assessment

arXiv:2610.09026v1 Announce Type: cross Abstract: We present BEACON-SP, an ontology-grounded Graph Retrieval-Augmented Generation (GraphRAG) framework for clinician-facing decision support in behavior

Beyond Reconstruction: What Matters in Action Tokenization for Robot Policies?

arXiv:2610.09170v1 Announce Type: new Abstract: Autoregressive action-token policies such as vision-language-action models require action tokenizers to translate discrete token sequences into precise

safetyarxiv-cs-ro

Tutorials

CERO: Where and When to Allocate Rollouts for RL Post-Training

arXiv:2610.09679v1 Announce Type: new Abstract: Adaptive rollout methods for group-relative reinforcement learning typically allocate a fixed per-update budget across prompts. We instead study how to

tutorialsarxiv-cs-lg

Certified by Abstention: Distribution-Free Guarantees for Chain-of-Thought Verifiers at Small Calibration Budgets

arXiv:2610.09541v1 Announce Type: cross Abstract: Signals that predict whether a chain-of-thought (CoT) trace is correct are compared by AUC, but deploying one requires a threshold with a guarantee. W

CM-DPO: Constraint-Margin Direct Preference Optimization for LLM Planning

arXiv:2610.09219v1 Announce Type: cross Abstract: Direct Preference Optimization (DPO) treats all constraint violations equally: a 1 budget overshoot and a 1,000 overshoot induce the same training sig

safetyarxiv-cs-lg