Search: “benchmark”

LLM Self-Correction with DeCRIM: Decompose, Critique, and Refine for Enhanced Following of Instructions with Multiple Constraints

arXiv:2410.06458v2 Announce Type: replace Abstract: Instruction following is a key capability for LLMs. However, recent studies have shown that LLMs often struggle with instructions containing multipl

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

MemSecBench: Tracking Agent Memory Poisoning from Persistence to Consequence and Repair

DGX agent

arXiv:2607.27080v1 Announce Type: cross Abstract: Memory systems allow agents to retain and reuse information from past interactions, but they can also let malicious content persist. A malicious instr

model-releasesarxiv-cs-ai

31 Jul 2026

Agents

Multi-Agent Debate Strategies: Survey, Taxonomy, and Challenges

arXiv:2607.26212v1 Announce Type: cross Abstract: Multi-Agent Debate (MAD) is a promising paradigm for improving the accuracy and robustness of Large Language Model (LLM)-based agentic systems. It ena

agentsarxiv-cs-ai

Hardware

ObjectStream: Latent Objects as Memory Anchors for Streaming Video Understanding

arXiv:2607.28312v1 Announce Type: new Abstract: Streaming video understanding requires models to continuously retain useful visual evidence before future questions are known. Existing approaches prima

hardwarearxiv-cs-cv

Safety

OPLD: On-Policy Latent Distillation for Multimodal Reasoning

arXiv:2607.28154v1 Announce Type: new Abstract: Interleaved multimodal Chain-of-Thought (CoT) improves visual reasoning by incorporating auxiliary visual evidence into intermediate reasoning. However,

safetyarxiv-cs-cv

OSReward: Instituting Standardized Evaluation for Cross-Platform Computer-Use Reward Models

arXiv:2607.28609v1 Announce Type: cross Abstract: Computer-using agents (CUAs) are advancing rapidly across the digital world. A CUA trajectory records the agent's actions, states, and reasoning. Veri

PAC-MAN: Perception-Aware CBF-RL for Whole-Body Safety in Humanoid Dodgeball

model-releasesarxiv-cs-ro

arXiv:2607.28623v1 Announce Type: new Abstract: We present PAC-MAN, a perception-aware CBF-RL framework that couples control-barrier safety with deployment-realistic onboard sensing for whole-body hum

SemAnCorr: Semantic Anchored Correspondence for Zero-Shot Manipulation Skill Transfer

model-releasesarxiv-cs-ro

arXiv:2607.28382v1 Announce Type: new Abstract: Transferring manipulation skills across object instances that share functionality but differ in geometry remains a fundamental challenge in robot learni

Shared Symbolic Backbones for Physically Consistent Multi-Output Symbolic Regression

arXiv:2607.26528v1 Announce Type: cross Abstract: Symbolic regression provides analytical expressions, but it is usually applied one output at a time. This is limiting in process systems, where state

Towards Generalized Synapse Detection Across Invertebrate Species

arXiv:2509.17041v2 Announce Type: replace Abstract: Behavioural differences across organisms, whether healthy or pathological, are closely tied to the structure of their neural circuits. Yet, the fine

Tycho: Active Abstraction with Programmatic World Models for ARC-AGI-3

arXiv:2607.28287v1 Announce Type: cross Abstract: ARC-AGI-3 turns abstraction into an interactive problem of skill acquisition. A player must infer an unfamiliar game's rules, hidden state, and goal w

UrbanDS: A Graph-Guided LLM Multi-Agent System for Data-Intensive Urban Tasks

arXiv:2607.26724v1 Announce Type: new Abstract: Large language model (LLM) agents have been widely applied in automating data science tasks. However, existing methods typically rely on a limited set o

Research

VETO: Towards Protecting Images From Frontier AI Editing

arXiv:2607.27292v1 Announce Type: new Abstract: The rise of powerful, accessible image-editing models such as FLUX.2 has brought high-fidelity editing within broad reach. Their capabilities now extend

researcharxiv-cs-cv

Research

ViP-Rig: Visual-Prompted Controllable Rigging

arXiv:2607.27982v1 Announce Type: new Abstract: Rigging is inherently task-dependent because the same mesh may require different skeletons and deformation behaviors across animation tasks. In practice

researcharxiv-cs-cv

Aligning LLM-Simulated and Human Examinees for Psychometric Calibration: A Cognitive Diagnostic Profiling Approach

arXiv:2607.26317v1 Announce Type: cross Abstract: Psychometric calibration for educational tests typically requires costly human response data. Large language models (LLMs) simulated examinees offer a

Budget-Aware LLM Discovery via Cost-Calibrated Frontier Utility

model-releasesarxiv-cs-lg

arXiv:2607.26828v1 Announce Type: new Abstract: Large language models increasingly support scientific and algorithmic discovery through inference-time search over evaluated candidates. Existing adapti

CMT-RAG: Complementary Memory Traces for Multi-turn Multi-hop RAG

arXiv:2607.26470v1 Announce Type: new Abstract: Multi-turn information-seeking conversations require both multi-hop reasoning and long-range dependency tracking across turns. However, existing RAG sys

DenseOn with the LateOn: Fully Open Dense and Late-Interaction Models for Multilingual, Long-Context, and Code Search

arXiv:2607.27178v1 Announce Type: new Abstract: State-of-the-art retrieval models increasingly rely on closed training data, creating a reproducibility gap. We present an open end-to-end recipe for tr

DialectLLM: A Dialect-Aware Dialog[ue] Generation Framework Beyond Standard American English

arXiv:2601.22888v4 Announce Type: replace Abstract: More than 80% of the 1.6B English speakers do not use Standard American English (SAE), yet LLMs often fail to correctly identify non-SAE dialects an

Safety

Do Methods Support the Claims? Intra-Paper Verification for Peer Review

arXiv:2607.26066v1 Announce Type: new Abstract: The growing volume of scientific submissions has motivated interest in using large language models (LLMs) to assist peer review. Existing automated nove

safetyarxiv-cs-cl

Safety

Do Unified Multimodal Models Think in One Space? A Lens Through Cross-Branch Steering

arXiv:2607.26411v1 Announce Type: new Abstract: Unified multimodal models (UMMs) aim to integrate understanding and generation within a single architecture, yet it remains unclear whether these capabi

safetyarxiv-cs-cv

Google DeepMind’s new AI model can control a robot’s entire body

model-releasesthe-verge-ai

Google DeepMind says the latest version of its Gemini Robotics AI model can 'control entire humanoid robots.' While the previous model focused on controlling a humanoid robot's upper body, Gemini Robo

How close are we to local llama robotics for consumer price point?

model-releasesr-localllama

I'm guessing 3 years, what do you think? In other words: many of us will be able to afford a general purpose robot in 3 years to experiment with in the home. Cost roughly $5k? Probably small size, but

ICDAR 2026 Competition on Information Extraction from Atomic Layer Deposition/Etching (ALD/E) Scientific Figures

arXiv:2607.26848v1 Announce Type: new Abstract: Scientific figure comprehension and reasoning using multimodal AI requires integrating visual perception with domain-specific reasoning to extract meani

Investigating reservoir computing for branch predictionin pipelined processors using emerging CMOS memristor devices

model-releasesarxiv-cs-lg

arXiv:2607.27140v1 Announce Type: cross Abstract: This project aimed to develop a novel reservoir compute (RC) implementation framework targeting high-speed operation and integration with CMOS digital

OmniAD: Detect and Understand Industrial Anomaly via Multimodal Reasoning

arXiv:2505.22039v2 Announce Type: replace Abstract: While anomaly detection has made significant progress, generating detailed analyses that incorporate industrial knowledge remains a challenge. To ad

Agents

Self-Configurable Mesh-Networks for Scalable Distributed Submodular Bandit Optimization

arXiv:2602.19366v2 Announce Type: replace-cross Abstract: We study how to scale distributed bandit submodular coordination under realistic communication constraints in bandwidth, data rate, and connec

agentsarxiv-cs-ro

Simultaneous Coverage and Efficiency Guarantee in Online Conformal Prediction

model-releasesarxiv-cs-lg

arXiv:2607.26577v1 Announce Type: new Abstract: Adaptive conformal inference (ACI) of Gibbs and Cand{es and its variants are the standard approach to online conformal prediction under distribution shi

TPCD: Tone-Pressure Contrastive Decoding and the Label-Free Gating Bottleneck in Vision-Language Models

arXiv:2607.26536v1 Announce Type: new Abstract: High-pressure prompts can push vision-language models (VLMs) into unsupported commitments, such as reading illegible text, reporting indeterminate times

VEGA: Learning Navigation VLAs from In-the-Wild Egocentric Video with Geometric Trajectory Supervision

model-releasesarxiv-cs-ro

arXiv:2606.18426v2 Announce Type: replace Abstract: We introduce VEGA, an approach for training navigation VisionLanguage-Action (VLA) models from unlabeled egocentric navigation videos. Internet-scal

Agent Retrieval Bench: Evaluating Repository Context Retrieval for Coding Agents

arXiv:2607.24882v1 Announce Type: cross Abstract: Modern coding agents are usually evaluated by whether they eventually produce a correct patch, but patch generation depends on an earlier context-acqu

AVE-Compass: Towards Holistic Evaluation for Audio-Video Editing Abilities

arXiv:2607.24821v1 Announce Type: cross Abstract: While instruction-based video editing has advanced rapidly, real-world videos contain tightly coupled audio and visual signals, and editing one modali

Built and released BetterGPT-150M – A compact 150M parameter completion model (+ live HF Space demo)

model-releasesr-localllama

Hey everyone, I recently finished pre-training BetterGPT-150M, a small, lightweight causal language model with ~152 million parameters.Trained on 15B tokens. Dataset & Training: Trained across stable

CLBench-V: Evaluating Multimodal Context Learning from Grounding to Knowledge Acquisition

arXiv:2607.25294v1 Announce Type: cross Abstract: Real-world tasks often require models to learn from task-specific context rather than relying only on pre-trained knowledge. While recent work has hig

CoTinyVLA: Chain-of-Thought Distillation for a Sub-Billion-Parameter Vision-Language-Action Model

arXiv:2607.25487v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models translate natural-language commands into robot action sequences, but leading systems on the LIBERO-Plus robustness b

COVENANT: Natural-Language Workflow Compilation for Aligned Agent Execution

arXiv:2607.25400v1 Announce Type: new Abstract: Large language model (LLM) agents are increasingly entrusted with natural-language workflow instructions (e.g., retail-payment policies) that specify no

Emergent Latent-State Computation under Stochastic Volatility

arXiv:2607.25459v1 Announce Type: cross Abstract: Mechanistic interpretability has largely focused on language models and deterministic toy tasks. Much less is known about how sequence models internal

Everyone is unique: Towards Behaviorally Heterogeneous Negotiation Dialogue Systems for Debt Collection

arXiv:2607.25218v1 Announce Type: new Abstract: Debt collection is a critical negotiation task in the financial industry, with strong practical relevance and exceptional academic value as a behavioral

From CUDA to MLX: How K-Search Brings Decades of Kernel Expertise to Apple Silicon

model-releasesberkeley-ai-research

Figure 1: CUDA-to-MLX optimization translation map. CUDA optimization knowledge can be translated into architecture-native MLX strategies rather than copied instruction-for-instruction. We face a new

Research

FunnelAL: Retrieve-then-Rank Active Learning for Single-Class Discovery

arXiv:2607.25276v1 Announce Type: new Abstract: We present FunnelAL, a retrieve-then-rank active learning system for single-class discovery, which adapts the multi-stage funnel architecture of industr

researcharxiv-cs-cv

I tried running a 1.56TB MoE model on a 6GB RTX 4050 Laptop, Here’s the result

model-releasesr-localllama

The Test Bench Setup I tested running a massive 1.56TB Mixture-of-Experts (MoE) checkpoint (96 shards, 93 layers, 896 experts/layer, ~4.46 bits/param MXFP4) on a budget gaming laptop. Laptop: HP Victu

Instruction-based Image Editing: A Survey on Data, Models, Evaluation, and Applications

arXiv:2607.25642v1 Announce Type: cross Abstract: Instruction-based Image Editing (IIE) aims to transform a given image into a new one based on textual instructions. Advances in Large Language Models

Kernel Forge: An Agent Harness for LLM-based Generation and Optimization of CUDA Kernels

arXiv:2607.24762v1 Announce Type: new Abstract: Machine learning models are increasingly embedded in everyday software, and most of their runtime is spent in a small set of compute kernels such as mat

MEDit-Bench: A Dataset for Evaluating Message-Driven Narrative Video Editing

arXiv:2607.25300v1 Announce Type: new Abstract: Video editing is fundamentally message-driven: even from the same source footage, the selected shots change depending on the narrative the editor wishes

OrchBench: Evaluating Multi-Agent Orchestration Plans in Isolation via Deterministic Simulation

arXiv:2607.25656v1 Announce Type: new Abstract: Complex tasks often decompose into parallelizable yet interdependent subtasks, making orchestration critical to the performance of multi-agent systems (

Pass the Baton: Trajectory-Relayed On-Policy Distillation

arXiv:2607.26057v1 Announce Type: cross Abstract: On-policy distillation (OPD) grounds token-level supervision in the student's own trajectory, yet suffers from prefix failure: once the student commit

Physics-Grounded Fluid Video Generation with a Simulation Dataset and Dual-Stream Optical-Flow Supervision

arXiv:2607.25321v1 Announce Type: new Abstract: Video diffusion models generate visually compelling content but routinely violate elementary physics when the subject involves fluids: liquid columns br

Polistemics: Evaluating LLMs as Information Mediators in Politics & Elections

arXiv:2607.25953v1 Announce Type: new Abstract: As LLMs increasingly mediate the political information citizens rely on, there is still no standardized way to assess whether they do so responsibly. We