Search: “benchmark”

Policy Alignment: New Signals for Membership Auditing in On-Policy Distillation

arXiv:2610.11423v1 Announce Type: new Abstract: On-policy distillation (OPD) trains a student model by aligning its policy with a teacher model on trajectories generated by the student model itself. T

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

Predicting Alignment Generalization with Value Representations

DGX agent

arXiv:2610.12410v1 Announce Type: cross Abstract: LLM developers post-train their models to exhibit prosocial values and behavioral traits, which are enumerated in an alignment target. However, while

model-releasesarxiv-cs-ai

9 Oct 2026

ProtoSemImage: Image-Valued Prototypes with Deformable Row Alignment for Interpretable Document Classification

arXiv:2610.11460v1 Announce Type: cross Abstract: Prototypes in classification models are almost always vectors, and a vector has no readable form. This paper asks what happens when a prototype is an

Local Ai

Qwen3.8 27b with 200K ctx + MTP on 12GB Ampere Cards

Hello all! Thank you to everyone that contributed their feedback and notes for LlamAmpere the last time I posted. I've continued to chip away at improvements for the 3090 crowd -- this release is mode

local-air-localllama

Randomized Transport Maps for Model-Free Policy-Gradient Mean-Field Control

arXiv:2610.11619v1 Announce Type: cross Abstract: We develop a model-free policy gradient method for discrete-time mean-field control (MFC). In MFC, the policy affects the objective both through the c

safetyarxiv-cs-lg

RaReCache: Bridging the Gap in Cross-Model KV Cache Reuse via Rank disagreement-based Selective Recomputation

arXiv:2610.11358v1 Announce Type: new Abstract: Cross-model KV-cache reuse remains a key challenge in modern LLM serving. Coding agents and multi-model systems increasingly route a shared context acro

REACT: Rolling Denoising and Dual Decoupling for Reactive Robot Control with VLA Models

arXiv:2610.12007v1 Announce Type: cross Abstract: Flow-based vision-language-action (VLA) models generate action chunks for temporally coherent robot motion, but chunked control creates a fundamental

Real Long-Term Memory for AI: A 50-Million-Token Window That Is Faster and Cheaper Than Recompute

arXiv:2610.10845v1 Announce Type: cross Abstract: A large language model can only use the text that fits in its context window, and it recomputes its internal key-value (KV) state for a prompt every t

Real-Time Motion Planning with Dynamic Hazards: Classical vs. Learning-Based Methods

arXiv:2610.12249v1 Announce Type: cross Abstract: We study real-time motion planning in dynamic hazard fields through a controlled comparison between classical planning and learning-based methods. Rat

Agents

ReCast: Attribution-Oriented Step Representation Learning for LLM-Based Agent Systems

arXiv:2610.11334v1 Announce Type: new Abstract: In LLM-based agent systems, failures can originate from early steps whose effects propagate through subsequent interactions, making their origins diffic

agentsarxiv-cs-ai

Recurrent Self-Improvement: Dynamic Cross-Loop On-Policy Distillation for Looped Language Models

arXiv:2610.10623v1 Announce Type: cross Abstract: Looped Language Models (LoopLMs) offer a parameter efficient approach to scaling reasoning by reusing shared parameters across recurrent computation s

Agents

Recursive Self-Improvement through Multi-Agent Self-Supervision

arXiv:2610.12176v1 Announce Type: new Abstract: Recursive self-improvement (RSI) of a model on non-verifiable tasks, such as open-ended research, faces a supervision bottleneck when its outputs exceed

agentsarxiv-cs-ai

Agents

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

Relative Patch Response Learning for Generalizable AI-Generated Image Detection

arXiv:2610.11876v1 Announce Type: new Abstract: Generative models can now synthesize highly realistic images, simultaneously increasing the risks of misinformation and visual forgery. Therefore, detec

safetyarxiv-cs-cv

REMORY: Learning Residual Memory for Context Compaction

arXiv:2610.11287v1 Announce Type: cross Abstract: Long-horizon agents compact their history to continue within a finite context window, but a textual summary alone may not support every subsequent dec

ReSI: Recursive Safety Improvement toward Resistant and Resilient AI

arXiv:2610.12233v1 Announce Type: cross Abstract: Recursive self-improvement, the participation of AI systems in improving their own capabilities, is beginning to move from theoretical prospect to pra

Residual Advantage: Student-Relative Teacher Guidance for RL with Verifiable Rewards

arXiv:2610.11519v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) and on-policy distillation (OPD) have become two main paradigms for post-training reasoning models

safetyarxiv-cs-cl

Residual Modeling Closes the Regression and Generative Policy Gap in Robot Learning

arXiv:2610.12231v1 Announce Type: new Abstract: Learning from demonstration has enabled impressive robot behaviors. A common choice for policy learning is to use diffusion or flow matching (Flow-Polic

safetyarxiv-cs-ro

ReTeach: Building a Self-Teacher through Multi-Round Reflection and Retry

arXiv:2610.11529v1 Announce Type: new Abstract: Self-distillation can improve reasoning without a separately trained, more capable teacher, but its effectiveness depends on how the self-teacher gains

safetyarxiv-cs-ai

Rethinking Contrastive Loss in CLIP Post-training: A Complementary Framework with Frozen Text Encoder

arXiv:2610.11374v1 Announce Type: cross Abstract: CLIP serves as a foundational vision-language model and the de facto vision encoder for downstream VLMs such as LLaVA. Post-training offers a lightwei

researcharxiv-cs-ai

Applications

Rewiring Semantics, Dynamics, and Control: A Simple yet Effective Action-Centric Tri-Stream Transformer

arXiv:2610.11416v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models have emerged as a prominent framework for complex robotic manipulation, building on the strong semantic understand

applicationsarxiv-cs-ai

RiCo: Neural Simulation of Rigid-Body Interactions via Local Contact Reasoning

arXiv:2610.12333v1 Announce Type: cross Abstract: Accurate simulation of rigid-body interactions is essential for predictive physical world models. Despite recent progress in modeling object dynamics,

RIFT: Relative Isolation From Trees For Anomaly Detection

arXiv:2610.12244v1 Announce Type: new Abstract: Isolation Forest (IF) is a widely used baseline for unsupervised anomaly detection. Recent studies provide a closed-form expression for the infinite-for

S4oP: Operator-level Pruning of Structured State Space Models for Resource-Constrained Devices

arXiv:2606.18096v2 Announce Type: replace-cross Abstract: Structured State Space Models (SSMs), including the S4 and S4D architectures, have recently emerged as powerful alternatives to attention-base

Safe at One Loop, Risky at Another: Aligning Safety Across Recurrent Depths in Looped Language Models

arXiv:2610.10625v1 Announce Type: cross Abstract: Looped Language Models (LoopLMs) provide a parameter efficient approach to scaling model capabilities through repeated use of shared parameters across

Safe, Persistent, and Evolving Agent Harness for Understanding Partially Observable Worlds

arXiv:2610.11552v1 Announce Type: new Abstract: Large language model agents can invoke tools fluently, but enterprise workflows demand more than selecting the right tools: actions must strictly comply

SAGE: Sink-Aware Guided Emphasis for Visual Grounding in Vision-Language Decoders

arXiv:2610.11469v1 Announce Type: cross Abstract: Recent large vision-language models (VLMs) pair a visual encoder with a large language model (LLM) and perform well on diverse image-text tasks, yet t

safetyarxiv-cs-cl

Sample-Efficiency of Kolmogorov-Arnold Networks

arXiv:2610.10627v1 Announce Type: new Abstract: Deep reinforcement learning has achieved substantial performance gains over classical control approaches. Yet, a central challenge to learning in real-w

SciTBERT: A family of chronologically consistent language models for scientific and technological language processing

arXiv:2610.12207v1 Announce Type: new Abstract: Pre-trained transformer models are increasingly being used to study scientific and technological progress. Encoders tuned to paper or patent text outper

SDPAD: A Fully Spike-Driven Pipeline for End-to-End Autonomous Driving

arXiv:2610.11583v1 Announce Type: cross Abstract: End-to-end autonomous driving demands trajectory planners that are both highly accurate and cheap enough for edge deployment. State-of-the-art artific

Seek-and-View Reasoning for Multi-View Spatial Understanding

arXiv:2610.11810v1 Announce Type: new Abstract: Existing approaches to multi-view spatial reasoning operate largely on sparse input views. Vision-language models (VLMs) are thus restricted to understa

safetyarxiv-cs-cv

Selective Listening: Mechanism-Guided Control of Audio Influence in Large Audio-Language Models

arXiv:2610.11196v1 Announce Type: cross Abstract: Large audio-language models (LALMs) exploit multimodal evidence, yet task-irrelevant audio can alter text-reasoning decisions when listening is unnece

researcharxiv-cs-cl

Self-Supervised Keyframe Discovery for Horizon-Invariant Behavior Cloning

arXiv:2610.10857v1 Announce Type: new Abstract: Behavior cloning (BC) in non-Markovian environments is a challenging problem because policies have to reason over contextual information over long horiz

Sera: Semantic Representation Aggregation for Reliable and Interpretable Battery Health Forecasting

arXiv:2610.11567v1 Announce Type: cross Abstract: Battery state of health (SoH) forecasting is important for battery management, but remains challenging due to nonlinear degradation and heterogeneity

SFT-as-Context Mitigates Forgetting in Supervised Fine-Tuning

arXiv:2610.11132v1 Announce Type: cross Abstract: Supervised fine-tuning (SFT) equips large language models (LLMs) with specialized capabilities, but often comes at the cost of forgetting the general

researcharxiv-cs-ai

SignRAG: Unified Retrieval-Augmented Gloss-Free Sign Language Translation

arXiv:2610.11371v1 Announce Type: new Abstract: Contemporary decoder-only large language models (LLMs) have demonstrated strong capabilities across a wide range of domains. However, existing pretraini

Slot3R: Set-Associative Spatial Memory for Streaming 3D Reconstruction

arXiv:2610.12282v1 Announce Type: new Abstract: Streaming 3D reconstruction must preserve evidence from each frame while processing an expanding scene online. Spatial memory is a natural fit because i

researcharxiv-cs-cv

SLVR: Structured Latent Visual Reasoning via Human-like Reasoning Flows

arXiv:2610.10563v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) often answer visual reasoning questions by relying on linguistic priors rather than task-relevant visual evid

researcharxiv-cs-ai

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

Sparse Attention Is Matrix Approximation, Not Choosing from a Bag of Values

arXiv:2610.10871v1 Announce Type: new Abstract: Large Language Models (LLMs) achieve strong performance across many domains, but their efficiency is limited by the quadratic cost of attention with res

researcharxiv-cs-cl

SparseDecoding: Decoding-Aware Pruning for Accurate and Efficient LLM Inference

arXiv:2610.12327v1 Announce Type: cross Abstract: The memory-bound nature of the decoding stage of large language model (LLM) inference incurs significant latency. Layer-wise training-free network pru

SpatialOPSD: Self-Distilling Spatial Intelligence from Verified Coding Agent Traces

arXiv:2610.11366v1 Announce Type: new Abstract: Spatial coding agents significantly improve spatial reasoning in Multimodal Large Language Models (MLLMs) by using external tools to generate verified e

safetyarxiv-cs-lg

SPD-MetaFormer is what you need for small-data brain decoding

arXiv:2610.10952v1 Announce Type: new Abstract: Brain signal decoding is challenging because neural recordings are noisy and vary across individuals, while labeled data are often limited. Recent atten

researcharxiv-cs-lg

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

STAG: A Sparse Traversability-Aware Graph Representation from Grid-Based Costmaps for Robotic Navigation

model-releasesarxiv-cs-ro

arXiv:2610.11943v1 Announce Type: new Abstract: Autonomous rovers navigating large unstructured environments need efficient global planning that accounts for terrain traversability. However, searching

SteerCast: Retrieval-Based Latent Steering for Decoder-Only Time Series Forecasting

arXiv:2610.11229v1 Announce Type: new Abstract: Time series forecasting aims to predict future values from historical observations and auxiliary features. We propose extbf{SteerCast}, a retrieval-base

researcharxiv-cs-lg

Applications

Stochastic Grouping Conformal Prediction for Effective Subgroup Reliability

arXiv:2610.11957v1 Announce Type: cross Abstract: Conformal prediction offers a distribution-free coverage guarantee, making it especially attractive for clinical applications. Standard conformal pred

applicationsarxiv-cs-ai

Stochastic Teacher Intervention for Agentic On-Policy Distillation

arXiv:2610.10878v1 Announce Type: cross Abstract: On-policy distillation (OPD) efficiently transfers capabilities from a stronger teacher to a student language model through dense token-level supervis