Search: “llm”

MotionBlind: Probing the Illusion of Motion Understanding in Video-LLMs

model-releasesarxiv-cs-cv

arXiv:2609.09528v1 Announce Type: new Abstract: Video large language models (Video-LLMs) are increasingly used as the perceptual front end of world models, a role that assumes they can read motion: ho

10 Sept 2026

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

SRD-GUARD: A Defense Framework of LLMs via Semantic Rewriting and Joint Multi-Model Scoring for Latent Intent Exposure

DGX agent

arXiv:2609.06540v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed in safety-critical applications, yet jailbreak attacks can conceal harmful intent through role-

model-releasesarxiv-cs-ai

10 Sept 2026

Why Is Video Still So Expensive? A Survey of Inference-Efficiency Mechanisms in Video and Audiovisual LLMs

arXiv:2609.10355v1 Announce Type: cross Abstract: Video understanding has rapidly evolved toward video large language models (VideoLLMs): systems that couple video representations with pretrained larg

10 Sept 2026

CausalBN-Bench: A Comprehensive Benchmark for Causal Learning Capability of LLMs

model-releasesarxiv-cs-lg

arXiv:2404.06349v3 Announce Type: replace Abstract: The ability to understand causality significantly impacts the competence of large language models (LLMs) in output explanation and counterfactual re

9 Sept 2026

You Can't Prefer Emotions You Don't Sample: Intensity Undershoot in DPO-Tuned LLMs

model-releasesarxiv-cs-lg

arXiv:2609.07808v1 Announce Type: cross Abstract: Ask a language model to respond 'very excitedly,' and its output is typically only mildly more energetic. We quantify this effect. We condition an ins

9 Sept 2026

I am seeing claims around LLMs, specifically Astra having made serious progress on robotics. The tasks that are demonstrated are simple pick…

I am seeing claims around LLMs, specifically Astra having made serious progress on robotics. The tasks that are demonstrated are simple pick and place tasks with parallel jaw grippers. LLMs can do pla

researchyann-lecun--x

8 Sept 2026

GUT: Quantifying and Optimizing the Reasoning Uncertainty of LLMs via Graph Complexity

arXiv:2609.05284v1 Announce Type: new Abstract: Recent years have witnessed great advances in the reasoning ability of Large Language Models (LLMs). However, the reasoning processes of LLMs often exhi

researcharxiv-cs-ai

7 Sept 2026

This dismissal of AI skepticism is egregiously wrong, or at least deeply selective, @jeffclune. Was I wrong that pure LLMs would always be v…

This dismissal of AI skepticism is egregiously wrong, or at least deeply selective, @jeffclune. Was I wrong that pure LLMs would always be vulnerable to hallucinations? Was I wrong that pure LLMs woul

safetygary-marcus--x

7 Sept 2026

Can LLMs Extract Architectural Design Decisions from Source Code Commits? - A Preliminary Exploratory Study

arXiv:2609.03721v1 Announce Type: cross Abstract: Context: Architectural Design Decisions (ADDs) capture the rationale behind the structure and evolution of software systems but are rarely documented

Caught in the Story: Narrative Captivity in Multi-turn LLMs Conversation

arXiv:2609.03407v1 Announce Type: new Abstract: People increasingly turn to large language models (LLMs) for everyday advice, making ethically charged interpersonal problems a practical moral-advisory

This work introduces diffusion-augmented LLMs, a new class of models. They first suggest that speculative decoding needs a separate draft mo…

This work introduces diffusion-augmented LLMs, a new class of models. They first suggest that speculative decoding needs a separate draft model. Diffusion LLMs give up the quality of the model they re

agentsdair-ai--x

Two-Stage Reinforcement Learning for Sound and Adversarial Test Generation in Code LLMs

arXiv:2609.03955v1 Announce Type: new Abstract: Reinforcement learning (RL) has substantially advanced code generation with large language models (LLMs) through executable feedback. The feedback for c

Impressive tool to explore frontier AI capabilities. Martian's AI Frontier lets you compare 44 LLMs by measured cost, quality, and reliabili…

Impressive tool to explore frontier AI capabilities. Martian's AI Frontier lets you compare 44 LLMs by measured cost, quality, and reliability, then see how routing and repeated sampling change the fr

agentsdair-ai--x

3 Sept 2026

Learn from Whoever Is Right: Answer-Verified Multi-Teacher Distillation for Multi-Domain LLMs

arXiv:2609.02548v1 Announce Type: cross Abstract: Modern large language models (LLMs) rely on reinforcement learning to build strong capabilities in individual domains, but integrating those capabilit

3 Sept 2026

User Feedback Provides a Unique Signal that LLMs Can not Detect

arXiv:2609.02859v1 Announce Type: new Abstract: Harnessing naturally occurring feedback from user interactions offers a promising learning signal for Large Language Models (LLMs). However, recent stud

3 Sept 2026

Do Multimodal LLMs See Before They Read? Diagnosing Contextual Sycophancy

arXiv:2609.00067v1 Announce Type: cross Abstract: External text can override conflicting image evidence in multimodal large language models, a failure we call multimodal contextual sycophancy. We intr

From Rollouts to Recipes: Self-Contained Post-Training for LLMs

arXiv:2609.01422v1 Announce Type: new Abstract: Post-training large language models usually applies a single training recipe to all samples, even though the model's own rollouts reveal different sampl

TempCloze: Can Video-LLMs Identify the Missing Middle?

arXiv:2609.01515v1 Announce Type: cross Abstract: Temporal reasoning benchmarks for Video-LLMs are often mediated by language, leaving room for linguistic shortcuts from option wording, answer correla

Triple-Bottom-Line Sustainability of Language Models for Edge AI: A Comparison Between SLMs and Quantized LLMs

arXiv:2609.00665v1 Announce Type: new Abstract: Edge-AI model selection is commonly driven by one isolated metric - accuracy, latency, memory, energy, or safety, even though a deployable language mode

Beyond Surface Alignment: Grounding the Dynamics of Situational Understanding and Generative Control in LLMs

arXiv:2608.29610v1 Announce Type: new Abstract: The current alignment tuning paradigm for Large Language Models (LLMs) prioritizes surface-level behaviors -- fluency, safety, and tonal consistency. Wh

BiG-SURE - Bipartite Graph for Semantic Uncertainty and Reliability Estimation of LLMs

arXiv:2608.30646v1 Announce Type: cross Abstract: Reliable uncertainty estimation is a crucial requirement for deploying large language models (LLMs) and vision-language models (VLMs) in safety-critic

Bridging Semantics and Physics with Constrained LLMs for Safe and Trustworthy Robotic Manipulation

arXiv:2608.29379v1 Announce Type: new Abstract: A language-guided robot operating in a real kitchen must do more than produce a plan that appears correct. It must also execute that plan safely in clut

safetyarxiv-cs-ro

Evaluating the Capabilities of LLMs for Persuasive Dialogue

arXiv:2608.29738v1 Announce Type: new Abstract: Large language models (LLMs) can generate apparently highly persuasive text, but does sounding persuasive mean arguing well? We introduce extsc{Persuasi

agentsarxiv-cs-cl

EvoGenUI-Bench: Evaluating LLMs as Multi-Turn Generative UI Assistants

arXiv:2608.29387v1 Announce Type: new Abstract: Large language models can generate interactive web interfaces, but reliable generative UI requires maintaining an executable artifact as user requests e

How Identity and Opinion Shape Political Sycophancy in LLMs

arXiv:2608.29198v1 Announce Type: new Abstract: As Large Language Models (LLMs) increasingly encourage users to disclose personal profiles for tailored assistance, measuring their political alignment

LaMoC: Loss-Aware Modular Compression for LLMs

arXiv:2608.30226v1 Announce Type: new Abstract: Modular compression has enabled considerable parameter reduction in LLMs while preserving strong language understanding and downstream task accuracy. Ho

Local Ai

LLMs Interpret, Embeddings Organize, Graphs Emerge: Agent-Driven Compilation of Scientific Knowledge

arXiv:2608.29612v1 Announce Type: new Abstract: Sustained scientific work requires a knowledge substrate that carries interpretation across tasks and preserves paths to source evidence. We call this p

local-aiarxiv-cs-ai

Low-Resource Preference Adaptation of LLMs via Activation-Based Label Propagation

arXiv:2608.30902v1 Announce Type: new Abstract: Adapting large language models to user-specific preferences is often constrained by the cost of human annotation, making preference optimisation impract

Not the Same Protector: Deployment-Dependent Protective Intervention in LLMs

arXiv:2608.29136v1 Announce Type: cross Abstract: We ask whether a model protects a user in the same way when that user speaks rather than types. Using a single distress vignette---a physical injury o

S3Gym: Can LLMs Turn Self-Testing and Self-Judging into Self-Improvement?

arXiv:2608.31100v1 Announce Type: new Abstract: Large language models (LLMs) increasingly interact with external environments and accumulate substantial behavioral experience, yet existing agent bench

ScienceArena: Benchmarking LLMs on Latest Scientific Olympiad Competitions

arXiv:2608.30517v1 Announce Type: new Abstract: Benchmark saturation and data contamination increasingly obscure genuine scientific reasoning in frontier LLMs. We introduce extsc{ScienceArena}, an oly

Applications

Unknown Unknowns: Do Hidden Intentions in LLMs Evade Detection?

arXiv:2601.18552v2 Announce Type: replace Abstract: LLMs expand accessibility and provide wide-reaching access to information. Yet these interactions also create opportunities to embed subtle, goal-or

applicationsarxiv-cs-cl

VisLens: Single-Pass Interpretable Visual Search for Multimodal LLMs

arXiv:2608.30705v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) struggle with fine-grained Visual Search, the task of locating small or rare objects in high-resolution images.

agentsarxiv-cs-cv

Learning a Single Token to Replace Long System Prompts in LLMs

arXiv:2511.23271v2 Announce Type: replace Abstract: Long system prompts are widely used to steer Large Language Models (LLMs), but repeatedly processing them at inference time is inefficient and consu

31 Aug 2026

Semantic Head Specialization Guides Hybrid ViT Attention for Multimodal LLMs

arXiv:2608.28383v1 Announce Type: cross Abstract: Hybrid attention dominates frontier LLMs, yet Vision Transformers (ViTs) in multimodal LLMs lack a satisfactory hybrid design, with no consensus on wh

31 Aug 2026

Applications

The Calls are Coming from Inside the Model: Investigating Probe-based Detection of Tool-Calling Errors in LLMs

arXiv:2608.27750v1 Announce Type: cross Abstract: The hidden states of large language models (LLMs) are known to capture rich information relating to model knowledge and behavior that can be hard to e

applicationsarxiv-cs-cl

31 Aug 2026

A Multi-Framework Comparison of Outline Stages in Long-Form Generation with LLMs

arXiv:2608.26177v1 Announce Type: cross Abstract: Long-form generation exposes fundamental limitations of large language models. Even 70B-parameter models exhibit length collapse at 16k-token outputs,

Activation Outliers Matter: Robust Recovery for Quantized Multimodal LLMs

arXiv:2608.26581v1 Announce Type: new Abstract: Low-bit quantization offers a promising avenue for reducing the computational and memory demands of Multimodal Large Language Models (MLLMs). Recent har

researcharxiv-cs-lg

Do LLMs Understand Personality? Rethinking Persona Fidelity Evaluation through Structured Behavioral Inference

arXiv:2608.26674v1 Announce Type: cross Abstract: As large language models are increasingly deployed to simulate diverse human characters, ensuring persona fidelity, defined as the extent to which an

agentsarxiv-cs-ai

From Atomic to Agentic: Towards Interpretable Evaluation of LLMs' Agentic Mathematical Capabilities

arXiv:2608.26950v1 Announce Type: new Abstract: Large Language Models (LLMs) are evolving from performing end-to-end mathematical reasoning to integrating agentic intelligence. However, most existing

Reason in the Words You Speak: Idiolectal Paraphrasing Off-Policy Traces for Reasoning Distillation in VideoLLMs

arXiv:2608.26684v1 Announce Type: new Abstract: Recent large language models achieve strong performance on complex reasoning tasks, where reinforcement learning with Group Relative Policy Optimization

safetyarxiv-cs-cv

Which India Survives Translation? Narrative Homogenisation Across Indian Oral Traditions in LLMs

arXiv:2608.26123v1 Announce Type: new Abstract: Large language models (LLMs) are trained predominantly on English-language internet text that over-represents certain cultural narratives, raising conce

Ban&Pick: Enhancing Performance and Efficiency of MoE-LLMs via Smarter Routing

model-releasesarxiv-cs-lg

arXiv:2509.06346v3 Announce Type: replace Abstract: Sparse Mixture-of-Experts (MoE) has become a key architecture for scaling large language models (LLMs) efficiently. Recent fine-grained MoE designs

From Specialization to Generalization: Instruction-tuned LLMs for Robust Harmful Content Mitigation

arXiv:2608.25605v1 Announce Type: new Abstract: Large language models (LLMs) demonstrate impressive performance across a wide range of general NLP tasks; however, their effectiveness in sensitive doma

Groundhog Bit-Flip Attack: Seeding Infinite Generation Loops in Mixture-of-Experts LLMs through Bit Flips

arXiv:2608.25276v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) architectures enable scalable and efficient large language models (LLMs) by selectively activating expert sub-networks through

agentsarxiv-cs-cl

Skill Issue: Are Skills Language-Invariant in LLMs?

arXiv:2608.25832v1 Announce Type: new Abstract: Large language models access knowledge inconsistently across languages, but to what extent do they differ in their skill sets when interacting with diff

Names Can Hurt: Spotting Slopsquatting Risks Caused by Package Name Hallucinations in Local Coding LLMs

arXiv:2608.23897v1 Announce Type: cross Abstract: When a code generating language model fabricates a Python package name, an adversary who has pre-registered that name on PyPI can convert that halluci

researcharxiv-cs-ai

26 Aug 2026

Can LLMs Truly Forget? Revealing Unlearning Gaps Through Adversarial Evaluation

arXiv:2608.21606v1 Announce Type: new Abstract: Machine unlearning aims to remove the influence of targeted training data from a model while preserving its remaining capabilities, but evaluating wheth