Search: “llm”

A Case Study on Emergent Cheating and Whistleblowing in Autonomous Research Swarms

arXiv:2609.04170v1 Announce Type: new Abstract: Multi-agent AI science ecosystems rely on agents possessing tools that allow them to communicate, coordinate, and build on each other's work. Yet this s

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

A computable representation of the physical laboratory enables verifiable workflows

DGX agent

arXiv:2609.03621v1 Announce Type: new Abstract: Making science computable requires representations of both scientific knowledge and the physical world in which scientific claims are tested. A computab

agentsarxiv-cs-ai

4 Sept 2026

Hardware

Achieving Extreme Efficiency through Specialized GPU Kernel Generation

This research details an agentic framework that automatically generates custom GPU kernels and rigorously validates them. By integrating automated synthesis and thorough performance testing, the syste

hardwaredatabricks

Building and Evaluating Fixed-Voice Thai TTS from Synthetic Speech

arXiv:2609.03502v1 Announce Type: cross Abstract: In low-resource settings, deploying TTS typically requires choosing between a large voice-cloning model with costly inference or a compact fixed-voice

CASCADE: A Component Ablation and Corpus Audit of a Layered Local Defense for MCP-Based Systems

arXiv:2604.17125v2 Announce Type: replace-cross Abstract: The Model Context Protocol (MCP) widens the prompt injection attack surface of large language model applications to tool descriptions, paramet

Causal-Counterfactual RAG: The Integration of Causal-Counterfactual Reasoning into RAG

arXiv:2509.14435v3 Announce Type: replace Abstract: Large language models (LLMs) have transformed natural language processing (NLP), enabling diverse applications by integrating large-scale pre-traine

Counterfactual Routing Using Integer Programming with Constraint Generation

arXiv:2609.03707v1 Announce Type: new Abstract: We present our submission to the IJCAI 2025 'Counterfactual Routing Competition' (CRC 25). The goal of the competition is to find counterfactual explana

researcharxiv-cs-ai

Crazy to see this. I have spent years saying that GenAI was unreliable and we can’t trust it. The fact that it cannot be trusted to follow i…

Crazy to see this. I have spent years saying that GenAI was unreliable and we can’t trust it. The fact that it cannot be trusted to follow instructions is a core part of the issue, as I have been poin

safetygary-marcus--x

Decoupling Turn-Taking from Semantics: A Decoupled Data Approach for Finite-State-Machine-Based Full-Duplex Dialogue

arXiv:2609.03321v1 Announce Type: new Abstract: The Neural Finite State Machine (NFSM) framework offers a pragmatic path to full-duplex dialogue by serializing turn-taking control and response generat

agentsarxiv-cs-cl

Efficient Test-Time Adaptation through Human-AI Interaction

arXiv:2609.04141v1 Announce Type: new Abstract: AI agents are trained on population-scale data to encode broad capabilities spanning those of many practitioners. Yet the artifacts they produce rarely

Environment Evolution for Terminal Agents

arXiv:2609.04128v1 Announce Type: new Abstract: Scaling interactive and verifiable environments is critical for training terminal agents. As frontier models become more capable, environments synthesiz

ESPO: Error-Structured Prompt Optimization via Diagnose, Diversify, and Stabilize

arXiv:2609.04197v1 Announce Type: cross Abstract: Evolutionary prompt optimizers such as GEPA suffer from prompt bloat: each iteration appends rules and caveats, producing prompts up to 3imes longer y

FlowBalance: Verifier-Grounded Self-Improvement from On-Policy Reasoning Experience

arXiv:2609.03241v1 Announce Type: cross Abstract: A reasoning model can improve from its own on-policy experience, but this inner loop is fragile: terminal verifiers provide reliable yet sparse superv

FLY-EVAL++: An Evidence-Driven Evaluation Protocol for Safety-Constrained Flight Prediction with Large Language Models

arXiv:2609.04021v1 Announce Type: new Abstract: Evaluating large language models (LLMs) in safety-critical, physics-governed environments requires more than accuracy-based metrics, because predictions

FPCO-Dialog: A Multi-Turn False-Premise Benchmark for Correction and Cooperation in Vision-Language Models

arXiv:2609.03331v1 Announce Type: new Abstract: Vision-language models (VLMs) are increasingly deployed in multi-turn settings where users may describe visual content with incorrect assumptions. Yet e

FrameBench:A Language Understanding Benchmark Based on Frame Semantics

arXiv:2609.03370v1 Announce Type: new Abstract: In frame semantics, sentence comprehension is assumed to proceed by relating lexical meaning to background knowledge called semantic frames, thereby ena

Free Pause Tokens

arXiv:2609.03807v1 Announce Type: cross Abstract: A free pause token gives a language model extra compute to form each next-token prediction (as a pause, or thinking, token does) but carries that comp

From Deceptive Outputs to Deceptive Mechanisms: A Causal Framework for Language-Model Deception Research

arXiv:2609.04166v1 Announce Type: new Abstract: Research and news coverage of language-model deception increasingly attributes human-like mental-state concepts to language models. Such claims can blur

researcharxiv-cs-ai

From Prior-Guided Heuristics to Deployable Agents: Accelerating Demonstration-Driven Reinforcement Learning for Deadline-Constrained Network Control

arXiv:2609.03590v1 Announce Type: cross Abstract: Timely delivery of delay-sensitive information over dynamic, heterogeneous networks is essential for NextG interactive applications, yet providing str

Good old days at OpenAI in 2016: an agent stares at screen pixels, moves a mouse, and books a flight on United. We called it World of Bits, …

Good old days at OpenAI in 2016: an agent stares at screen pixels, moves a mouse, and books a flight on United. We called it World of Bits, inside OpenAI Universe. 10 yrs later, Astra is reincarnated

safetyjim-fan--x

GPS-Bench: A Governance Policy Benchmark for Automating Policy Analysis

arXiv:2609.03553v1 Announce Type: new Abstract: Policy analysis requires more than predicting whether a proposal will pass: it requires identifying who will be affected, how those actors respond, and

Headroom-Drift Replay: A Primitive for Principled Replay Control in GRPO

arXiv:2609.03941v1 Announce Type: cross Abstract: RL-based post-training for reasoning models is increasingly bottlenecked by repeated fresh rollout generation, particularly in agentic settings where

Insightful paper from Microsoft and colleagues. If you have ever had an agent run fail 80 steps ago with no way to find where, this one is f…

Insightful paper from Microsoft and colleagues. If you have ever had an agent run fail 80 steps ago with no way to find where, this one is for you. (bookmark it) Agent failures show up as long complex

agentsdair-ai--x

Investigating the Ability of Large Language Models to Analyze Recipes for Diabetes

arXiv:2609.03967v1 Announce Type: cross Abstract: Several studies have evaluated the ability of Large Language Models (LLMs) for meal planning, yielding positive outcomes. These models can process nat

Language, Language Models, and What We're Talking About

arXiv:2609.03577v1 Announce Type: new Abstract: Language models are commonly discussed as technical artefacts, but they are obviously shaped by the linguistic worlds conveyed by data during their trai

Legibility is Not Interpretability: Comparing Judged and Actual Importance in Chain-Of-Thought Reasoning

arXiv:2609.04194v1 Announce Type: new Abstract: Reasoning traces from chain-of-thought models appear to offer a legible window into how a model arrives at its answer. A growing body of work treats the

Less Is Moral: A CHARMing Framework for Moral Foundations Detection in Endorsement Behaviour

arXiv:2609.03330v1 Announce Type: new Abstract: Moral language plays a central role in shaping online endorsement and the diffusion of information, yet existing moral foundation detection systems ofte

safetyarxiv-cs-cl

Lngram v2: Latent N-Gram Memory with Interpretable Discrete Representations

arXiv:2609.03426v1 Announce Type: new Abstract: Transformers lack a native lookup mechanism, requiring repeated dense computation to recognize and reuse local static patterns. Lngram v1 introduces tok

Measuring Harmfulness of Computer-Using Agents

arXiv:2508.00935v3 Announce Type: replace-cross Abstract: Computer-using agents (CUAs), which can autonomously control computers to perform multi-step actions, might pose significant safety risks if m

MedQA-MM: Shortcuts Behind Medical Visual Reasoning

arXiv:2609.03261v1 Announce Type: cross Abstract: A benchmark score credits final answers, but not the route by which an item can be answered. In medical multimodal multiple-choice questions (MCQs), t

ObserverBench: Testing Mechanistic Estimates for Intervention and Control

arXiv:2609.03026v1 Announce Type: cross Abstract: Mechanistic interpretability is increasingly used to guide interventions such as activation steering, circuit removal, and safety monitoring. Yet an i

PACE: Towards Surfacing Hidden Conflicts in User Requests

arXiv:2609.03293v1 Announce Type: new Abstract: Personalized assistants should not only comply with user requests but also assess whether those requests are appropriate given the user's current circum

safetyarxiv-cs-cl

R^{2}Adapter: A Routing and Rewriting Adapter for Efficient Hybrid RAG

arXiv:2609.02894v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) has become a prevailing paradigm for enhancing Large Language Models (LLMs) with non-parametric knowledge. Vanilla

Representational alignment yields generalizable safety in language models

arXiv:2609.04022v1 Announce Type: cross Abstract: Aligning large language models (LLMs) is essential for their safe deployment. Current alignment methods mainly optimize observable responses, yet mode

SafeRestore: Detector-Relative Risk Certificates for Selective Industrial Image Restoration

arXiv:2609.03475v1 Announce Type: new Abstract: Industrial inspection pipelines often restore a measured image before a detector acts on it, yet restoration can suppress detector-supported defect stru

safetyarxiv-cs-cv

SGD-KV: Summarization Guided KV Cache Compression

arXiv:2609.03235v1 Announce Type: new Abstract: Large language models (LLMs) face severe memory bottlenecks in long-context inference due to the linearly growing size of key-value (KV) caches. Existin

SWE-Gate: Passing Functional Tests Is Not Enough for Software Engineering Agents

arXiv:2609.04167v1 Announce Type: cross Abstract: Repository-level software engineering benchmarks have significantly advanced the evaluation of coding agents, but existing benchmarks primarily measur

The Pelican comparison grid for Astra is pretty interesting

model-releasessimon-willison

I got access to GPT-6 Astra this afternoon, so naturally I used it to generate SVGs of pelicans riding bicycles - at low, medium, high, xhigh and max reasoning levels (Astra doesn't support reasoning=

To What Extent Do Large Language Models Understand Bangla Idioms?

arXiv:2609.03410v1 Announce Type: new Abstract: Idiomatic expressions are an integral part of natural language, reflecting cultural nuances and posing unique challenges for computational models, parti

Towards Numerical TOHTN Planning with SMT-based HTN-SAT Encoding

arXiv:2609.03938v1 Announce Type: new Abstract: While HTN planning has received significant attention in recent years, support for numerical reasoning remains very limited. In this paper, we investiga

Transfiver: Human-AI Co-Inference through a Shared Editable State

arXiv:2609.03797v1 Announce Type: new Abstract: Long-term human-AI interaction is difficult because the information that guides inference is updated implicitly by the model and is not directly inspect

We open-sourced Paddock, our Rust/C++ inference engine with its own CUDA kernels (MIT/Apache-2.0)

model-releasesr-localllama

I'm one of the developers. We said in August it would go open source in September and it did last night. MIT or Apache-2.0, pick one. The repo you see is our internal repo, kernels included, so from n

When Models Edit Too Much: On the Fidelity of Minimal Code Edits

arXiv:2609.04061v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used to edit existing code, but correctness alone is not enough: useful repairs should also be minimal,

Action abstractions for amortized sampling

arXiv:2410.15184v2 Announce Type: replace-cross Abstract: As trajectories sampled by policies used by reinforcement learning (RL) and generative flow networks (GFlowNets) grow longer, credit assignmen

AI Mathematician: Towards Fully Automated Frontier Mathematical Research

arXiv:2505.22451v2 Announce Type: replace Abstract: Large Reasoning Models (LRMs) have made significant progress in mathematical capabilities in recent times. However, these successes have been primar

Before the Script, Set the Stage: How Worldview Simulation Amplifies Psychologically Grounded Persuasion in Multi-Turn Jailbreaking

arXiv:2609.02414v1 Announce Type: cross Abstract: Multi-turn jailbreak attacks demonstrate that harmful intent can be distributed across dialogue, yet existing methods obscure what conversational mech

Beyond-RAG: Question Identification and Answer Generation in Real-Time Conversations

arXiv:2410.10136v2 Announce Type: replace-cross Abstract: In customer contact centers, human agents often struggle with long average handling times (AHT) due to the need to manually interpret queries

Breadth Beats Depth: Improving GCG-Based Jailbreak Optimization with Breadth-Oriented Suffix Search

arXiv:2609.02172v1 Announce Type: new Abstract: Optimization-based jailbreak attacks such as Greedy Coordinate Gradient (GCG) achieve strong effectiveness and transferability by optimizing adversarial

safetyarxiv-cs-cl