Search: “reasoning”

When Self-Evolution Backfires: Pre-Commit Gating against Skill Contamination in LLM Agents

arXiv:2608.05810v1 Announce Type: new Abstract: Self-evolving agents accumulate capability by distilling reusable skills from their execution trajectories, but we find this process is not monotonic: p

7 Aug 2026

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

2 x 5070ti Qwen 27B full config / stats

DGX agent

Following up on yesterday's post about running everyone's faves on 2 x 16gb cards while maximizing performance and KV. Previous post data used abandoned Cu130 VLLM image. Stats here are done on cu129-

model-releasesr-localllama

6 Aug 2026

Anyone understand what the equivalent of GPT-5.6 Instant in ChatGPT is for the OpenAI API?

model-releasessimon-willison--x

The tweet is a question from user Simon Willison (posted on 6 Aug 2026) asking which OpenAI API model corresponds to the ChatGPT “GPT‑5.6 Instant” version. No answer or clarification is included in th

Auto-fit vs tuned MoE offload: 564 → 1330 pp tok/s, unchanged decode (Qwen3.6-35B-A3B Q6 / RTX 3090)

model-releasesr-localllama

TL;DR: On a Qwen3.6-35B-A3B Q6 setup sized for 64K context on a 24GB RTX 3090, spilling eight MoE expert layers to CPU freed enough VRAM to increase -b from 512 to 1024 and -ub from 128 to 512. Prompt

BridgeVLA++: A Data-Efficient, Generalizable, and Memory-Augmented Vision-Language-Action Framework for 3D Manipulation

arXiv:2608.05042v1 Announce Type: new Abstract: Leveraging pre-trained vision-language models (VLMs) to construct vision-language-action (VLA) models has emerged as a promising paradigm for 3D robot m

safetyarxiv-cs-ro

Caching for the Future: Scrub Jay Episodic Memory Principles for Agent Memory Systems

arXiv:2608.04746v1 Announce Type: new Abstract: LLM agents that persist across sessions accumulate stored memories whose validity varies enormously by content type, yet existing memory architectures t

CoCo-IR: Contextual Composed Image Retrieval

arXiv:2608.05149v1 Announce Type: new Abstract: Current instruction-based image retrieval systems are powerful but limited to single-turn interactions, failing to capture the iterative nature of compl

DeepAmbigQA: Ambiguous Multi-hop Questions for Benchmarking LLM Answer Completeness

researchapple-ml-research

Large language models (LLMs) with integrated search tools show strong promise in open-domain question answering (QA), yet they often struggle to produce complete answer set to complex questions such a

DSeq-JEPA: Discriminative Sequential Joint-Embedding Predictive Architecture

arXiv:2511.17354v4 Announce Type: replace Abstract: Recent advances in self-supervised visual representation learning have demonstrated the effectiveness of predictive latent-space objectives for lear

researcharxiv-cs-cv

Emergence of Hierarchical Emotion Organization in Large Language Models

arXiv:2507.10599v3 Announce Type: replace-cross Abstract: As large language models (LLMs) increasingly power conversational agents, understanding how they model users' emotional states is critical for

EvtGraph: Event-Adaptive Compression for Sparse Temporal Graph Learning in Multimodal Time Series

arXiv:2608.04368v1 Announce Type: new Abstract: Multimodal temporal data are inherently irregular and uneven in information density, yet most models rely on uniform discretization, leading to ineffici

researcharxiv-cs-lg

ExeCRE: Execution-Consistency Guided Reliability Estimation for Self-Correcting Code Generation

arXiv:2608.04439v1 Announce Type: cross Abstract: Large language models (LLMs) have made notable progress in code generation, but they still struggle on challenging tasks that require sophisticated al

Explicit Language Memory for Long-Horizon Planning in Vision-Language-Action Models

arXiv:2608.04765v1 Announce Type: cross Abstract: Vision-language-action (VLA) models provide a unified paradigm for connecting visual perception, language understanding, and robotic control. However,

Faster-WAM: Efficient Inference-Time Future Conditioning for Robust World Action Models

arXiv:2608.04404v1 Announce Type: new Abstract: World Action Models (WAMs) improve robot manipulation by learning how the environment evolves beyond the current observation. However, existing approach

From Brute Force to Semantic Insight: Performance-Guided Data Transformation Design with LLMs

arXiv:2601.03808v2 Announce Type: replace Abstract: Large language models (LLMs) have achieved notable performance in code synthesis; however, data-aware augmentation remains a limiting factor, handle

safetyarxiv-cs-cv

Governing Execution Risk in Agentic AI Systems: A Trajectory-Guided Framework for Red Teaming

arXiv:2608.04018v1 Announce Type: cross Abstract: AI agents are increasingly embedded in organizational workflows, where they interact with external information sources and invoke digital tools to per

safetyarxiv-cs-ai

Imitation Learning from Human Motion Alone Does Not Guarantee Biomechanically Plausible Gait Kinetics

arXiv:2603.12408v3 Announce Type: replace-cross Abstract: Motion imitation learning (IL) is increasingly used in robotics and human gait modeling, yet its ability to recover biomechanically consistent

researcharxiv-cs-lg

K-EXAONE 2.0 Technical Report

arXiv:2608.04505v1 Announce Type: new Abstract: This technical report presents K-EXAONE 2.0, an open-weight multilingual foundation model developed by LG AI Research as a step in our effort toward glo

KV cache quantization benchmarks: 413 pairs tested on Qwen 3.6 27B, Gemma 4 31B. KLD with BeeLlama.cpp v0.4.0: KVarN 6-bit beats q8_0, precision tail 1024 dominates

model-releasesr-localllama

Link to the article: KV Cache Quantization Benchmarks: KVarN, Precision Tail KLD benchmarks with BeeLlama.cpp v0.4.0, fork of llama.cpp with more KV cache quantization options. Models: Qwen 3.6 27B Q5

Mamba with Hierarchical Memory: Solving Representation Bottleneck in Long Sequence Modeling

arXiv:2608.02347v2 Announce Type: replace Abstract: Recurrent linear attention models (RLAs) such as Mamba offer efficient linear-time sequence modeling as an alternative to Transformers, yet their fi

Not Truly Multilingual: Script Consistency as a Missing Dimension in VLM Evaluation

arXiv:2606.17188v3 Announce Type: replace-cross Abstract: Current multilingual evaluations for Vision-Language Models (VLMs) assume a one-to-one mapping between language and orthography, overlooking b

OmniEdit-Bench: A Comprehensive Benchmark for Instruction-based Video Editing

arXiv:2608.05049v1 Announce Type: new Abstract: Instruction-based video editing (IVE) is an emerging field with broad applications, yet evaluating editing models remains challenging. Existing benchmar

PriDyG: Privacy-preserving Dynamic Graph Inference with LLM-GNN Collaboration

arXiv:2608.04255v1 Announce Type: cross Abstract: Graph inference over relational data can expose sensitive edge information, and this risk becomes more severe in dynamic graphs, where repeated model

researcharxiv-cs-lg

ResPlan: A Large-Scale Vector-Graph Dataset of 17,000 Residential Floor Plans

arXiv:2508.14006v2 Announce Type: replace Abstract: We introduce ResPlan, a dataset of 17,000 residential floor plans with vector geometry, room-connectivity graphs, and metric-scale coordinates. Each

Risk Is Not the Target: A Monotonic Framework for Evaluating Wildfire Operational Risk Signals

arXiv:2607.21597v2 Announce Type: replace Abstract: Evaluating wildfire risk systems using standard machine-learning metrics such as F1-score or IoU is fundamentally flawed: these metrics assess event

local-aiarxiv-cs-ai

SafeCommit: Certifying When Memory-Grounded Agents May Safely Act

arXiv:2608.04289v1 Announce Type: new Abstract: Long-horizon agents increasingly use persistent memory and tools to take actions with external side effects. A central failure mode is premature commitm

safetyarxiv-cs-ai

Tutorials

Spatiotemporal Graph Transformer for Traffic Intelligence in Edge Computing

arXiv:2608.04075v1 Announce Type: cross Abstract: Accurate traffic forecasting is essential for proactive resource management in edge computing, where service demand evolves dynamically across both sp

tutorialsarxiv-cs-ai

Agents

stratum: A System Infrastructure for Massive Agent-Centric ML Workloads

arXiv:2603.03589v3 Announce Type: replace-cross Abstract: Recent advances in large language models (LLMs) transform how machine learning (ML) pipelines are developed and evaluated. LLMs enable a new t

agentsarxiv-cs-lg

Tutorials

SurgNarrator: A Generative Retrieval Framework for Surgical Video Understanding

arXiv:2608.04676v1 Announce Type: new Abstract: Surgical procedures unfold as structured and recurring clinical events, whose real-time understanding via intraoperative surgical videos is critical for

tutorialsarxiv-cs-cv

SVI-DAG: A Structured Variational Inference Approach to Bayesian Causal Discovery

arXiv:2608.04930v1 Announce Type: cross Abstract: Bayesian causal discovery seeks to determine the posterior distribution of causal theories, which are interpreted as directed acyclic graphs (DAGs) th

Teaching MLLMs to Say No: Generalized Referring Expression Comprehension via Refusal Calibrated GRPO

arXiv:2608.04698v1 Announce Type: cross Abstract: We tackle the challenging yet underexplored task of Generalized Referring Expression Comprehension (GREC), which requires a model to localize the obje

local-aiarxiv-cs-ai

The Order Is the Guarantee: Verifier-Budgeted Code Deletion with Static-First Learned Proposals

arXiv:2608.04611v1 Announce Type: cross Abstract: Frontier coding models now match or exceed strong human reference points on programming benchmarks, yet benchmark success does not imply maintainable

The Yokai Learning Environment: Tracking Beliefs Over Space and Time

arXiv:2508.12480v3 Announce Type: replace Abstract: The ability to cooperate with unknown partners is a central challenge in cooperative AI and widely studied in the form of zero-shot coordination (ZS

Towards End-to-End Multilingual Metaphor Processing: Integrating Detection, Translation, and Evaluation

arXiv:2608.04260v1 Announce Type: new Abstract: Metaphorical language remains a major challenge for multilingual natural language processing because successful interpretation and translation require r

researcharxiv-cs-cl

Trident : How to Break Deep Reinforcement Learning Cyber Defenses (Agentic)

arXiv:2608.04317v1 Announce Type: cross Abstract: Autonomous cyber defense systems based on Deep Reinforcement Learning (DRL) have attracted significant research attention, yet remain evaluated almost

Uncertainty-aware Predict-Then-Optimize Framework for Equitable Post-Disaster Power Restoration

arXiv:2508.04780v2 Announce Type: replace-cross Abstract: The increasing frequency of extreme weather events, such as hurricanes, highlights the urgent need for efficient and equitable power system re

Tutorials

Unforgettable Generalization in Language Models

arXiv:2409.02228v2 Announce Type: replace-cross Abstract: When language models (LMs) are trained to forget (or 'unlearn'') a skill, how precisely does their behavior change? We study the behavior of t

tutorialsarxiv-cs-cl

Visual Representation Matters: Exploiting Temporal Differences in Video-to-Audio Generation

arXiv:2608.04902v1 Announce Type: new Abstract: Video-to-audio (V2A) generation extends image-to-audio generation (I2A) by introducing consecutive frames that provide essential temporal cues for audio

A Unified 2D Framework for DeepLesion Detection, Segmentation and Short Report Generation

arXiv:2608.02805v1 Announce Type: cross Abstract: In previous work, we integrated large language models (LLMs) into the lesion segmentation model based on the ULS23 DeepLesion dataset, using short-for

Agents

After rigorous testing, our joint AI project with Daiwa Securities is entering the full-scale production phase. We're bringing our agentic A…

After rigorous testing, our joint AI project with Daiwa Securities is entering the full-scale production phase. We're bringing our agentic AI systems to @Daiwa_JP’s wealth management teams to accelera

agentsdavid-ha--x

Agent memory layers don't need an LLM deciding what to remember

Most agent memory setups run a model call on the way in. Something reads the turn, decides whether it's worth keeping, rewrites it into a 'memory', tags it with a type and an importance score. That's

local-air-localllama

AgenticSCR: An Autonomous Agentic Secure Code Review for Immature Vulnerabilities Detection

arXiv:2601.19138v2 Announce Type: replace-cross Abstract: Secure code review is critical during pre-integration, where Atlassian developers rely on lightweight analysis tools, while deep security asse

local-aiarxiv-cs-ai

AI Security Leaderboard: Methodology, Results and Minimal Standard

arXiv:2608.03070v1 Announce Type: cross Abstract: Frontier AI model developers increasingly rely on layered safeguards to prevent catastrophic misuse, but little public evidence exists on how much pro

Aligned in Form, Not in Meaning: The Comprehension - Containment Decoupling of LLM Safety in Low-Resource Bangla Derogatory Speech

arXiv:2608.02941v1 Announce Type: new Abstract: We audit five frontier large language models on native Bangla derogatory speech (gali) across six protocols to test a single hypothesis: Comprehension-C

safetyarxiv-cs-cl

ANNOTARES: A Dataset for Extracting Logical Structures from German Statutory Texts

arXiv:2608.03898v1 Announce Type: new Abstract: The automatic structural analysis of legal texts is a cornerstone of legal technology, yet the extraction of their logical components remains a signific

Anyone interested in building a harness-only benchmark?

model-releasesr-localllama

There are a lot of LLM benchmarks but few, if any, harness benchmarks. I am thinking this would be a really good community project to build one. End goal: a leaderboard of harness performance (multipl

Agents

Autoreflection: How Agentic Strange Loops Turn Human Culture into AI Infrastructure

arXiv:2608.03800v1 Announce Type: cross Abstract: An LLM-based agent is a loop that reads itself. Agentic frameworks externalize identity, memory, and disposition into editable files. The agent loads

agentsarxiv-cs-ai

Applications

Benchmarking the Benchmarks: Testing the Predictive Validity of Commonsense Benchmarks

arXiv:2608.03340v1 Announce Type: new Abstract: Predicting LLM's capabilities on real-world tasks is essential, yet the extent to which performance on commonsense benchmarks predicts downstream perfor

applicationsarxiv-cs-cl