Search: “llm”

Allegory of the Cave: Measurement-Grounded Vision-Language Learning

arXiv:2605.11727v1 Announce Type: cross Abstract: Vision-language models typically reason over post-ISP RGB images, although RGB rendering can clip, suppress, or quantize sensor evidence before infere

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

AlphaGRPO: Unlocking Self-Reflective Multimodal Generation in UMMs via Decompositional Verifiable Reward

DGX agent

arXiv:2605.12495v1 Announce Type: new Abstract: In this paper, we propose AlphaGRPO, a novel framework that applies Group Relative Policy Optimization (GRPO) to AR-Diffusion Unified Multimodal Models

safetyarxiv-cs-cv

13 May 2026

Asymmetric Advantage Modulation Calibrates Entropy Dynamics in RLVR

arXiv:2604.04894v2 Announce Type: replace Abstract: Reinforcement learning with verifiable rewards (RLVR) has substantially improved the reasoning ability of large language models (LLMs), but it often

Local Ai

b9134

local-aillama-cpp-releases

B9134 is a recent build release of llama.cpp, an open-source C/C++ library for efficient large language model inference. The release was tagged on May 13, 2026, representing an ongoing development ite

Built a local coding harness powered by Gemma 4. It runs locally, connects to my model backend, starts coding sessions, streams responses, a…

Built a local coding harness powered by Gemma 4. It runs locally, connects to my model backend, starts coding sessions, streams responses, and uses tools through a CLI-style workflow. Still early, but

model-releasesollama--x

Hardware

Cerebras — Faster Tokens Please

Cerebras, a company specializing in AI accelerators and wafer-scale computing systems, is discussed in terms of its approaches to improving token generation speed in large language models, which is cr

hardwaresemianalysis

ClinicalBench: Stress-Testing Assertion-Aware Retrieval for Cross-Admission Clinical QA on MIMIC-IV

arXiv:2605.11143v1 Announce Type: new Abstract: Reasoning benchmarks measure clinical performance on clean inputs. We evaluate the step before reasoning: retrieval over real EHR notes, where negation,

Context Convergence Improves Answering Inferential Questions

arXiv:2605.12370v1 Announce Type: new Abstract: While Large Language Models (LLMs) are widely used in open-domain Question Answering (QA), their ability to handle inferential questions-where answers m

Covering Human Action Space for Computer Use: Data Synthesis and Benchmark

arXiv:2605.12501v1 Announce Type: new Abstract: Computer-use agents (CUAs) automate on-screen work, as illustrated by GPT-5.4 and Claude. Yet their reliability on complex, low-frequency interactions i

Agents

Deep Reasoning in General Purpose Agents via Structured Meta-Cognition

arXiv:2605.11388v1 Announce Type: new Abstract: Humans intuitively solve complex problems by flexibly shifting among reasoning modes: they plan, execute, revise intermediate goals, resolve ambiguity t

agentsarxiv-cs-cl

Don't Pass@k: A Bayesian Framework for Large Language Model Evaluation

arXiv:2510.04265v4 Announce Type: replace-cross Abstract: Pass@k is widely used to report the reasoning performance of LLMs, but it often produces unstable and potentially misleading rankings, especia

Enhancing Multilingual Counterfactual Generation through Alignment-as-Preference Optimization

arXiv:2605.11632v1 Announce Type: new Abstract: Self-generated counterfactual explanations (SCEs) are minimally modified inputs (minimality) generated by large language models (LLMs) that flip their o

EsoLang-Bench: Evaluating Genuine Reasoning in Large Language Models via Esoteric Programming Languages

arXiv:2603.09678v2 Announce Type: replace-cross Abstract: Large language models achieve near-ceiling performance on code generation benchmarks, yet most of the programming languages used by popular be

EvoNav: Evolutionary Reward Function Design for Robot Navigation with Large Language Models

arXiv:2605.11859v1 Announce Type: new Abstract: Robot navigation is a crucial task with applications to social robots in dynamic human environments. While Reinforcement Learning (RL) has shown great p

safetyarxiv-cs-ro

Fine-Tuning Large Language Models for Cooperative Tactical Deconfliction of Small Unmanned Aerial Systems

model-releasesarxiv-cs-ro

arXiv:2603.28561v2 Announce Type: replace Abstract: The growing deployment of small Unmanned Aerial Systems (sUASs) in low-altitude airspaces has increased the need for reliable tactical deconfliction

Applications

From Token to Token Pair: Efficient Prompt Compression for Large Language Models in Clinical Prediction

arXiv:2605.11774v1 Announce Type: new Abstract: By processing electronic health records (EHRs) as natural language sequences, large language models (LLMs) have shown potential in clinical prediction t

applicationsarxiv-cs-cl

Google Named a Leader in the Gartner® Magic Quadrant™ for AI Application Development Platforms: Mid-cycle update

model-releasesgoogle-cloud-ai

May 2026 update: We’ve refreshed this post to reflect our mid-cycle positioning and the evolution of our platform since the report was first published last November. Last fall, Google was recognized a

GRAFT-ATHENA: Self-Improving Agentic Teams for Autonomous Discovery and Evolutionary Numerical Algorithms

arXiv:2605.11117v1 Announce Type: new Abstract: Scientific discovery can be modeled as a sequence of probabilistic decisions that map physical problems to numerical solutions. Recent agentic AI system

gym-invmgmt: An Open Benchmarking Framework for Inventory Management Methods

arXiv:2605.11355v1 Announce Type: new Abstract: Inventory-policy comparisons are often difficult to interpret because performance depends on the evaluation contract as much as on the policy itself. Di

How far can bias go? Tracing bias from pretraining data to alignment

arXiv:2411.19240v2 Announce Type: replace Abstract: As LLMs are increasingly integrated into user-facing applications, addressing biases that perpetuate societal inequalities is crucial. While much wo

Human-Grounded Multimodal Benchmark with 900K-Scale Aggregated Student Response Distributions from Japan's National Assessment of Academic Ability

arXiv:2605.11663v1 Announce Type: new Abstract: Authentic school examinations provide a high-validity test bed for evaluating multimodal large language models (MLLMs), yet benchmarks grounded in Japan

Tutorials

i provided a clear road map in the essay Next Decade in AI in arxiv in 2020. not one person in the field has given a serious critique of it.…

i provided a clear road map in the essay Next Decade in AI in arxiv in 2020. not one person in the field has given a serious critique of it. if you are not aware of it, it is not my fault; i have poin

tutorialsgary-marcus--x

Interpreting Context-Aware Human Preferences for Multi-Objective Robot Navigation

arXiv:2603.17510v2 Announce Type: replace Abstract: Robots operating in human-shared environments must not only achieve task-level navigation objectives such as safety and efficiency, but also adapt t

safetyarxiv-cs-ro

Agents

Langsmith Engine appears to be a seriously high value release from the @LangChain team A deep agent that’s monitoring Langsmith agent traces…

Langsmith Engine appears to be a seriously high value release from the @LangChain team A deep agent that’s monitoring Langsmith agent traces and not only surfaces issues but: - proposes a PR to resolv

agentsharrison-chase--x

Large Language Models for Causal Relations Extraction in Social Media: A Validation Framework for Disaster Intelligence

arXiv:2605.11348v1 Announce Type: new Abstract: During disasters, extracting causal relations from social media can strengthen situational awareness by identifying factors linked to casualties, physic

Modeling Narrative Structure in Latin Epic Poetry with Automatically Generated Story Grammars

arXiv:2502.12276v2 Announce Type: replace Abstract: Computational methods for analyzing prose and poetry utilize word embeddings and other abstract representations that sometimes obscure context-rich

MotionBench: Benchmarking and Improving Fine-grained Video Motion Understanding for Vision Language Models

arXiv:2501.02955v2 Announce Type: replace Abstract: In recent years, vision language models (VLMs) have made significant advancements in video understanding. However, a crucial capability - fine-grain

Red-Teaming Text-to-Image Models via In-Context Experience Replay and Semantic-Preserving Prompt Rewriting

arXiv:2411.16769v3 Announce Type: replace-cross Abstract: Understanding the capabilities of text-to-image (T2I) models in harmful content generation is essential to safety and compliance. However, hum

Tutorials

Reflect then Learn: Active Prompting for Information Extraction Guided by Introspective Confusion

arXiv:2508.10036v2 Announce Type: replace Abstract: Large Language Models (LLMs) show remarkable potential for few-shot information extraction (IE), yet their performance is highly sensitive to the ch

tutorialsarxiv-cs-cl

RETUYT-INCO at BEA 2026 Shared Task 2: Meta-prompting in Rubric-based Scoring for German

arXiv:2605.11242v1 Announce Type: new Abstract: In this paper, we present the RETUYT-INCO participation at the BEA 2026 shared task 'Rubric-based Short Answer Scoring for German'. Our team participate

Rotation-Preserving Supervised Fine-Tuning

arXiv:2605.10973v1 Announce Type: new Abstract: Supervised fine-tuning (SFT) improves in-domain performance but can degrade out-of-domain (OOD) generalization. Prior work suggests that this degradatio

researcharxiv-cs-lg

Applications

Self-Consistent Latent Reasoning: Long Latent Sequence Reasoning for Vision-Language Model

arXiv:2605.12163v1 Announce Type: new Abstract: In language reasoning, longer chains of thought consistently yield better performance, which naturally suggests that visual latent reasoning may likewis

applicationsarxiv-cs-cv

Support-Proximity Augmented Diffusion Estimation for Offline Black-Box Optimization

arXiv:2605.11246v1 Announce Type: new Abstract: Offline black-box optimization aims to discover novel designs with high property scores using only a static dataset, a task fundamentally challenged by

Test-Time Compute for Dense Retrieval: Agentic Program Generation with Frozen Embedding Models

arXiv:2605.11374v1 Announce Type: cross Abstract: Test-time compute is widely believed to benefit only large reasoning models. We show it also helps small embedding models. Most modern embedding check

The Challenge and Reward of Fair Play in Narrative: A Computational Approach

arXiv:2507.13841v2 Announce Type: replace Abstract: Good storytelling involves surprise -- unpredictability in how the story unfolds -- and sense-making, the requirement that the story forms a coheren

Very Efficient Listwise Multimodal Reranking for Long Documents

arXiv:2605.11864v1 Announce Type: cross Abstract: Listwise reranking is a key yet computationally expensive component in vision-centric retrieval and multimodal retrieval-augmented generation (M-RAG)

We built SmithDB: the database purpose built for agent observability workloads that now powers many parts of LangSmith. Agent observability …

model-releasesharrison-chase--x

We built SmithDB: the database purpose built for agent observability workloads that now powers many parts of LangSmith. Agent observability presents a challenging data problem. Agent traces can contai

When Emotion Becomes Trigger: Emotion-style dynamic Backdoor Attack Parasitising Large Language Models

arXiv:2605.11612v1 Announce Type: new Abstract: Backdoor vulnerabilities widely exist in the fine-tuning of large language models(LLMs). Most backdoor poisoning methods operate mainly at the token lev

When to Ask a Question: Understanding Communication Strategies in Generative AI Tools

arXiv:2605.11240v1 Announce Type: cross Abstract: Generative AI models differ from traditional machine learning tools in that they allow users to provide as much or as little information as they choos

safetyarxiv-cs-lg

3DReflecNet: A Large-Scale Dataset for 3D Reconstruction of Reflective, Transparent, and Low-Texture Objects

arXiv:2605.10204v1 Announce Type: new Abstract: Accurate 3D reconstruction of objects with reflective, transparent, or low-texture surfaces still remains notoriously challenging. Such materials often

Tutorials

A Semantic-Sampling Framework for Evaluating Calibration in Open-Ended Question Answering

arXiv:2605.08432v1 Announce Type: cross Abstract: Calibration measures whether a model's predicted confidence aligns with its empirical accuracy, and is central to the reliable deployment of large lan

tutorialsarxiv-cs-ai

A Single Layer to Explain Them All:Understanding Massive Activations in Large Language Models

arXiv:2605.08504v1 Announce Type: new Abstract: We investigate the origins of massive activations in large language models (LLMs) and identify a specific layer named the extbf{Massive Emergence Layer

Local Ai

Accelerating Zeroth-Order Spectral Optimization with Partial Orthogonalization from Power Iteration

arXiv:2605.09034v1 Announce Type: new Abstract: Zeroth-order (ZO) optimization has become increasingly popular and important in fine-tuning large language models (LLMs), especially on edge devices due

local-aiarxiv-cs-lg

Agent-ValueBench: A Comprehensive Benchmark for Evaluating Agent Values

model-releasesarxiv-cs-ai

arXiv:2605.10365v1 Announce Type: new Abstract: Autonomous agents have rapidly matured as task executors and seen widespread deployment via harnesses such as OpenClaw. Safety concerns have rightly dra

AIPO: : Learning to Reason from Active Interaction

arXiv:2605.08401v1 Announce Type: cross Abstract: Recent advances in large language models (LLMs) have demonstrated remarkable reasoning capabilities, largely stimulated by Reinforcement Learning with

safetyarxiv-cs-ai

Ambig-DS: A Benchmark for Task-Framing Ambiguity in Data-Science Agents

model-releasesarxiv-cs-ai

arXiv:2605.09698v1 Announce Type: new Abstract: As data-science agents shift from co-pilots to auto-pilots, silent misframing becomes a critical failure mode. Agents quietly commit to plausible but un

An Annotation Scheme and Classifier for Personal Facts in Dialogue

arXiv:2605.10339v1 Announce Type: new Abstract: The advancement of Large Language Models (LLMs) has enabled their application in personalized dialogue systems. We present an extended annotation scheme

Arcane: An Assertion Reduction Framework through Semantic Clustering and MCTS-Guided Rule Exploring

model-releasesarxiv-cs-ai

arXiv:2605.10107v1 Announce Type: new Abstract: Assertion-based Verification (ABV) is essential for ensuring that hardware designs conform to their intended specifications. However, existing automated