Search: “models”

Token-Level Generalization in LoRA Adapter Backdoors: Attack Characterization and Behavioral Detection

arXiv:2605.30189v1 Announce Type: cross Abstract: We show that LoRA adapters, the dominant distribution format for fine-tuned LLMs, can be reliably backdoored through training data poisoning while pre

29 May 2026

Model Releases

Toward Ethical Facial Age Estimation: A Generalized Zero-Shot Benchmark Without Training on Children's Data

DGX agent

arXiv:2605.29230v1 Announce Type: cross Abstract: Age estimation from facial images typically relies on training data that includes images of minors, a practice that raises serious ethical, legal, and

model-releasesarxiv-cs-ai

29 May 2026

Training Deliberative Monitors for Black-Box Scheming Detection

arXiv:2605.29601v1 Announce Type: cross Abstract: As autonomous agents become more capable of performing real-world tasks, distinguishing scheming behavior from benign task pursuit may become a centra

29 May 2026

Who can we trust? LLM-as-a-jury for Comparative Assessment

arXiv:2602.16610v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly applied as automatic evaluators for natural language generation assessment often using pairwise

29 May 2026

A Query Engine for the Agents

arXiv:2605.27785v1 Announce Type: new Abstract: The fastest-growing data in production today is unstructured text: agent traces, chat logs, reasoning chains, model outputs. People want to analyze it,

AsyncTool: Evaluating the Asynchronous Function Calling Capability under Multi-Task Scenarios

arXiv:2605.27995v1 Announce Type: new Abstract: Large language model (LLM)-based agents have shown strong capabilities in using external tools to solve complex tasks. However, existing evaluations oft

Automating Formal Verification with Agent-Guided Tree Search

model-releasesarxiv-cs-lg

arXiv:2605.27485v1 Announce Type: cross Abstract: Formal verification offers a path to provably correct software, but writing verified code remains expensive enough that the technique is rarely used i

Behavioural Analysis of Alignment Faking

arXiv:2605.27681v1 Announce Type: new Abstract: Alignment faking (AF) refers to a model strategically complying with a training objective to avoid behavioural modification while preserving its deploym

safetyarxiv-cs-ai

Bounded-Compute Multimodal Regression for Product-Rating Prediction

model-releasesarxiv-cs-cv

arXiv:2605.27737v1 Announce Type: new Abstract: Vision-language models (VLMs) are increasingly attractive for multimodal quality assessment, but their default reliance on autoregressive text generatio

Confidence-Orchestrated Self-Evolution against Uncertain LLM Feedback

arXiv:2605.28010v1 Announce Type: new Abstract: Self-evolving large language models (LLMs) learn by generating their own training tasks and solutions, reducing reliance on human-curated supervision. H

Data-Efficient On-Policy Distillation for Automatic Speech Recognition

arXiv:2605.28139v1 Announce Type: new Abstract: Building competitive automatic speech recognition (ASR) models usually requires large-scale au- dio supervision, which makes reproduction and specializa

Deepfake-Eval-2024: A Multi-Modal In-the-Wild Benchmark of Deepfakes Circulated in 2024

arXiv:2503.02857v5 Announce Type: replace-cross Abstract: In the age of increasingly realistic generative AI, robust deepfake detection is essential for mitigating fraud and disinformation. While many

DEPART: DEcomposing PARiTy across Multilingual LLMs

arXiv:2605.28163v1 Announce Type: cross Abstract: Multilingual Large Language Models (mLLMs) leaderboards report per-language accuracy but rarely explain why disparities emerge, leaving systemic biase

Do LLMs Favor Their Providers? Measuring Vertical Integration Bias in Code Generation

arXiv:2605.28515v1 Announce Type: cross Abstract: Large Language Models (LLMs) have become an integral part of software development, especially with the advent of agentic capabilities. Yet, many front

EvoSpec: Evolving Speculative Decoding via Real-Time Vocabulary and Parameter AdaptationTarget

arXiv:2605.27390v1 Announce Type: cross Abstract: Speculative decoding accelerates Large Language Model inference via a draft-then-verify paradigm, yet the output projection layer becomes a bottleneck

FABSVer: Faster Training and Better Self-Verification for LLM Mathematical Reasoning

arXiv:2605.28389v1 Announce Type: new Abstract: While large language models have made significant progress in mathematical reasoning, they remain unreliable at judging the correctness of their own sol

safetyarxiv-cs-cl

Framing Matters: Addressing Framing Sensitivity in Decision-Making through Behaviorally-Grounded Value Alignment

arXiv:2605.28188v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly deployed in high-stakes decision-making settings such as legal reasoning, where consistency under factuall

Functional Entropy: Predicting Functional Correctness in LLM-Generated Code with Uncertainty Quantification

arXiv:2605.28500v1 Announce Type: cross Abstract: Large language models have shown impressive capabilities in code generation, yet they often produce functionally incorrect code. Uncertainty quantific

KSAFE-MM: A Multimodal Safety Benchmark via Localized Contextualization for Korean Cultural Risks

arXiv:2605.28013v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) exacerbate safety risks by introducing vulnerabilities across multiple modalities, such as language and vision.

LACUNA: Safe Agents as Recursive Program Holes

arXiv:2605.28617v1 Announce Type: new Abstract: LLM agents increasingly act by writing code, yet a split persists between the runtime that drives the agent and the code the model writes. The runtime o

safetyarxiv-cs-ai

Tutorials

Learn from your own latents and not from tokens: A sample-complexity theory

arXiv:2605.27734v1 Announce Type: new Abstract: Generative models, from diffusion models to large language models, achieve remarkable performance but at a cost in training data orders of magnitude lar

tutorialsarxiv-cs-lg

llm-anthropic 0.25.1

model-releasessimon-willison

Release: llm-anthropic 0.25.1 New model: Claude Opus 4.8 (claude-opus-4.8). New -o fast 1 option for fast mode, for organizations with that feature enabled on their account. Default max_tokens for eac

ProgVLA: Progress-Aware Robot Manipulation Skill Learning

model-releasesarxiv-cs-lg

arXiv:2605.28231v1 Announce Type: cross Abstract: We present ProgVLA, a compact vision-language-action (VLA) model designed for reliable robot manipulation under tight compute and memory budgets. The

SNARE: Adaptive Scenario Synthesis for Eliciting Overeager Behavior in Coding Agents

arXiv:2605.28122v1 Announce Type: cross Abstract: A coding agent executes a benign task as a sequence of shell, file, and network actions, any of which can quietly exceed the authorized scope while th

SuperValid: Capability-Aligned OOD Validation for Generalizable Downstream Scaling

arXiv:2605.28179v1 Announce Type: new Abstract: Scaling laws guide large language model training by relating compute to cross-entropy loss, and recent work further extends them to predict downstream b

Tabero: Learning Gentle Manipulation with Closed-Loop Force Feedback from Vision, Touch, and Language

model-releasesarxiv-cs-ro

arXiv:2605.27886v1 Announce Type: new Abstract: Tactile sensing is essential for robots to achieve human-like gentle manipulation. However, existing Vision-Language-Action (VLA) models struggle to exp

Research

The Well-Tempered Classifier: Some Elementary Properties of Temperature Scaling

arXiv:2602.14862v2 Announce Type: replace-cross Abstract: Temperature scaling is a simple method that allows to control the uncertainty of probabilistic models. It is mostly used in two contexts: impr

researcharxiv-cs-ai

UniMaia: Steering Chess Policies with Language for Human-like Play

arXiv:2605.27767v1 Announce Type: cross Abstract: Recent advances in large language models have enabled natural language to serve as a flexible interface for controlling complex systems, but often at

When do complex-valued neural networks help? A study of representation, geometry, and optimization

model-releasesarxiv-cs-lg

arXiv:2605.27673v1 Announce Type: new Abstract: Complex-valued Neural Networks (CVNNs) are often motivated by domains where information is naturally encoded in magnitude and phase. Yet complex-valued

ZipRL: Adaptive Multi-Turn Context Compression with Hindsight Response Replay

arXiv:2605.28069v1 Announce Type: new Abstract: Adaptive context compression is vital for scaling Large Language Models (LLMs) to complex, multi-turn agent tasks. However, rule-based compression metho

A Hybrid Vision-Language Architecture for Automated Defect Reasoning and Report Generation in Industrial Inspection

arXiv:2605.26533v1 Announce Type: cross Abstract: Automated industrial inspection requires both precise defect localization and structured maintenance report generation; in current practice these task

A Universal Cliff and a Design Fingerprint: Cross-Section Defect Detection Under LLM Orchestration

arXiv:2605.26174v1 Announce Type: cross Abstract: Production language-model systems answer a request by partitioning it across an invisible orchestration of worker agents that recompose one integrated

safetyarxiv-cs-ai

Beyond a Single Direction: Chain-of-Thought Disrupts Simple Steering of Refusal

arXiv:2605.26772v1 Announce Type: new Abstract: Large reasoning models (LRMs) generate chain-of-thought (CoT) traces before producing final outputs, introducing a dynamic internal state that may compl

Research

Beyond Semantics: The Unreasonable Effectiveness of Reasonless Intermediate Tokens

arXiv:2505.13775v4 Announce Type: replace-cross Abstract: Recent impressive results from large reasoning models have been interpreted as a triumph of Chain of Thought (CoT), and especially of the proc

researcharxiv-cs-ai

ChartAct: A Benchmark for Dynamic Chart Understanding

model-releasesarxiv-cs-cv

arXiv:2605.26994v1 Announce Type: new Abstract: Charts are widely used to present complex data for analysis and decision making. Existing chart understanding benchmarks mainly focus on static charts,

ConVer: Using Contracts and Loop Invariant Synthesis for Scalable Formal Software Verification

arXiv:2605.27051v1 Announce Type: cross Abstract: Formal verification of large C programs is impeded by state-space explosion: Bounded Model Checking (BMC) tools must encode the entire state space up

Tutorials

Demystifying Video Reasoning

arXiv:2603.16870v2 Announce Type: replace-cross Abstract: Recent advances in video generation have revealed an unexpected phenomenon: diffusion-based video models exhibit non-trivial reasoning capabil

tutorialsarxiv-cs-ai

Entropy Sentinel: Continuous LLM Accuracy Monitoring from Decoding Entropy Traces in STEM

arXiv:2601.09001v4 Announce Type: replace Abstract: Deploying LLMs raises two coupled challenges: (1) monitoring--estimating where a model underperforms as traffic and domains drift--and (2) improveme

'Give Me BF16 or Give Me Death'? Accuracy-Performance Trade-Offs in LLM Quantization

arXiv:2411.02355v4 Announce Type: replace-cross Abstract: Quantization is a powerful tool for accelerating large language model (LLM) inference, but the accuracy-performance trade-offs across differen

GraphDancer: Training LLMs to Explore and Reason over Graphs via Two-Stage Curriculum Post-Training

arXiv:2602.02518v2 Announce Type: replace-cross Abstract: Large language models (LLMs) increasingly rely on external knowledge to improve factuality, yet many real-world knowledge sources are organize

Memory Architectures for Multi-Turn Text-to-SQL: A Benchmark and Empirical Study

arXiv:2605.26394v1 Announce Type: new Abstract: Multi-turn Text-to-SQL is central to enterprise analytics yet remains predominantly evaluated in single-turn settings. We introduce EnterpriseMem-Bench,

Research

Once-For-All: A Train-Once and Select-Anytime Framework for Multimodal Instruction Tuning

arXiv:2605.26761v1 Announce Type: new Abstract: Multimodal instruction tuning is the de facto recipe for adapting vision language models (VLMs), yet instruction data are highly redundant, making data

researcharxiv-cs-cv

PersLitEval: Fine-grained Benchmark and Evaluation of LLMs on Persian Literature Questions

arXiv:2605.27015v1 Announce Type: new Abstract: Despite impressive multilingual capabilities, large language models (LLMs) remain poorly evaluated on literary knowledge in non-English languages. We in

Qiskit QuantumKatas: Adapting Microsoft's Quantum Computing exercises for LLM evaluation

arXiv:2605.27210v1 Announce Type: cross Abstract: We adapt Microsoft's QuantumKatas -- a well-established quantum computing curriculum -- from Q# to Qiskit, the most widely-adopted quantum computing f

ReMoE: Boosting Expert Reuse through Router Fine-Tuning in Memory-Constrained MoE LLM Inference

arXiv:2605.27081v1 Announce Type: cross Abstract: Fine-grained Mixture-of-Experts (MoE) models sparsely activate only a subset of experts per token, reducing activated computation while maintaining hi

Representation-Aware Unlearning via Activation Signatures: From Suppression to Entity-Signature Erasure

arXiv:2601.10566v5 Announce Type: replace Abstract: Entity-level unlearning is usually evaluated by what a model says: whether it stops naming the target, refuses a query, or shifts a Truth Ratio dist

Rethinking the Multilingual Reasoning Gap with Layer Swap

arXiv:2605.26735v1 Announce Type: new Abstract: Recent reasoning Large Language Models produce a chain-of-thought (CoT) predominantly in English, even when prompted in non-English languages. Prior wor

Separating Semantic Competition from Context Length in RAG Reading

arXiv:2605.27294v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) systems can respond incorrectly even when the correct passage was retrieved. The model must still read the retrieve