Search: “evaluation”

Explore, Then Commit: Measurement-Efficient Scientific Law Discovery with Language Models

arXiv:2610.07620v1 Announce Type: new Abstract: Scientific law discovery requires selecting measurements and converting evidence into a governing equation. We evaluate an explore-then-commit protocol

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Agents

Fork-and-Flush: Escaping Idea Basins in Autoresearch Agents

DGX agent

arXiv:2610.07447v1 Announce Type: new Abstract: Autoresearch agents tackle open-ended problems by repeatedly proposing candidate solutions, evaluating them, and using feedback to guide subsequent expe

agentsarxiv-cs-lg

7 Oct 2026

From Evidence to Action: How Tool-Using Agents Fail

arXiv:2610.07753v1 Announce Type: cross Abstract: Tool-using agents make consequential changes to external state, yet correct outcomes do not guarantee that their actions were supported by evidence es

researcharxiv-cs-ai

Grounding What Shapes the Plan: Rethinking Groundedness for Physical Intelligence in Autonomous Driving

arXiv:2610.07521v1 Announce Type: new Abstract: Driving models increasingly ground reasoning in causal relations, spatial structure, perceptual evidence, and predicted futures. These advances make rea

Harmful SFT Leaves a Continuous Trace in LLM Checkpoint Updates

arXiv:2610.07518v1 Announce Type: cross Abstract: Safety auditing of post-trained large language models typically relies on model behavior, requiring model execution and depending on the coverage of a

How Many Independent Samples Does a Satellite Image Contain? Generalization Bounds for Spatially Dependent Data

arXiv:2610.08227v1 Announce Type: cross Abstract: Machine learning classifiers for remote sensing imagery are typically evaluated as though every pixel were an independent sample. Spatial autocorrelat

Agents

How much of AutoResearch is research, and how much is search?[D]

I've recently been working part-time on an AutoResearch-style project. The setup is roughly: humans take recent work from top-tier ML/AI conferences, turn part of it into a well-defined task with an e

agentsr-machinelearning

Jailbreaking Open-Weight LLMs via Random Embedding Perturbations

arXiv:2610.07125v1 Announce Type: cross Abstract: While open-weight models have enjoyed steady progress in capabilities and wide adoption across multiple domains, their safety remains an important con

Less Is More: A Leakage-Controlled Study of Dermoscopic Preprocessing for Joint Skin Lesion Classification and Segmentation with YOLO26

arXiv:2610.08570v1 Announce Type: cross Abstract: Handcrafted preprocessing is widely employed in automated dermoscopic analysis to suppress imaging artifacts and enhance lesion visibility. Neverthele

Machine Learning for German Redispatch Forecasting under Data Delays and Temporal Distribution Shift

arXiv:2610.08337v1 Announce Type: cross Abstract: Public redispatch records provide empirical data for grid congestion forecasting, but delayed reporting, zero-inflated distributions, and temporal shi

MedCORE: Criteria-Grounded Clinical Reasoning for Interpretable Medical Image Diagnosis

arXiv:2610.08528v1 Announce Type: cross Abstract: Clinical diagnosis is inherently a structured reasoning process, yet existing deep learning models often bypass this structure by mapping image featur

researcharxiv-cs-ai

MiniScope: Authorizing Agents with Least-Privilege Permissions

arXiv:2512.11147v2 Announce Type: replace-cross Abstract: AI agents are increasingly granted autonomous access to sensitive user data and third-party services, making effective permission management a

Monocular Navigation Relative to Unknown Spacecraft Using a Transformer-Aided Kalman Filter

arXiv:2610.07231v1 Announce Type: cross Abstract: This work presents a novel learning-based pipeline for pose estimation of unknown spacecraft using only monocular images from a single servicer. The a

Local Ai

Neural Fields Encode Adaptation Geometry

arXiv:2610.07253v1 Announce Type: new Abstract: Neural fields are usually evaluated by how well they reconstruct an observation. We show that this misses two useful properties of a fitted network: how

local-aiarxiv-cs-lg

Offline AI Modules: Voice-First Offline Architecture, Hardware Reference Stack, Quantization and Benchmarking

arXiv:2610.07026v1 Announce Type: new Abstract: The Offline AI Modules workstream enables practical, low-power, and community-accessible deployment of voice-first AI systems that operate fully offline

ParanoiaEval: Benchmarking Unnecessary Defensive Work in Agentic Coding

arXiv:2610.08662v1 Announce Type: new Abstract: As coding agents increasingly undertake real-world work autonomously, judging whether their risk treatments are warranted has become important. Existing

PDB: Point-Based Deformation Blending for Facial Animation Retargeting

arXiv:2610.08672v1 Announce Type: new Abstract: Mesh-agnostic facial animation retargeting transfers expressions across meshes with different structures, but preserving facial motion without surface a

Protective Perturbations Must Survive the Resize: Scale-Robust Image Immunization against Malicious Editing

arXiv:2610.07464v1 Announce Type: new Abstract: Protective perturbations aim to stop malicious instruction-guided editing of personal photos, but they are optimized and evaluated at the editor's worki

Reading, Not Manipulating: Leveraging Router Logits for Multimodal Safety in MoE Vision-Language Models

model-releasesarxiv-cs-cl

arXiv:2610.07774v1 Announce Type: new Abstract: Vision-language models (VLMs) face compositional safety risks where harmful intent emerges from the interaction between visual and textual inputs. As mi

SAFESHIELD: A Decision-Organization Framework for Deployment-Time Safety of Small Language Models

arXiv:2610.07276v1 Announce Type: cross Abstract: Deployment-time safety of language models is commonly implemented through runtime guardrails such as input moderation, routing, retrieval verification

SalArt-VQA: Diagnosing Whether VLMs Understand Salient Artifacts in Generated Images

model-releasesarxiv-cs-cv

arXiv:2606.12671v2 Announce Type: replace Abstract: Vision-language models (VLMs) are increasingly used to detect whether AI-generated images contain visible artifacts, yet their ability to analyze su

Same Feedback, Different Answer: Measuring Run-to-Run Instability in Frontier-Model Customer Feedback Analysis

arXiv:2610.08036v1 Announce Type: new Abstract: AI agents are increasingly being programmed to automate knowledge work over large collections of unstructured data. Such automation requires repeatabili

Same-Number Citation Swaps: Stress-Testing Jev as a Financial Evidence Judge

model-releasesarxiv-cs-cl

arXiv:2610.08675v1 Announce Type: new Abstract: Financial reports repeat values across periods, metrics and accounting lines, allowing an LLM-generated calculation to be numerically correct while citi

Applications

Scalable extraction and visualization of multi-attribute logical and functional dependencies in tabular data

arXiv:2610.08287v1 Announce Type: new Abstract: Understanding the structural relationships among attributes in tabular data is fundamental to machine learning and pattern recognition. While functional

applicationsarxiv-cs-lg

ServeLearnBench: How Well Can Agents Self-Improve from Serving Experience?

arXiv:2610.07792v1 Announce Type: cross Abstract: Large language model agents are increasingly deployed to perform complex tasks in real-world environments. However, the knowledge required for correct

SIGMA: Self-Improving Alignment Generalization from a Model Spec

arXiv:2610.07935v1 Announce Type: new Abstract: LLM agents are increasingly capable of executing complex tasks and of recursively improving themselves on easy-to-verify objectives such as software eng

Agents

Stateless Language Agents: Scaling Long-Horizon Automated Research

arXiv:2610.07625v1 Announce Type: cross Abstract: Automated research systems increasingly run LLM agents over long horizons, but more inference does not by itself produce more progress: agents replay

agentsarxiv-cs-ai

Surviving the Router: Optimizing Skill Injections for Retrieval and Execution

arXiv:2610.08098v1 Announce Type: cross Abstract: AI agents increasingly rely on modular third-party 'skills' that are dynamically selected by skill routers to execute complex tasks. While recent stud

TabiBERT: A Large-Scale ModernBERT Foundation Model and A Unified Benchmark for Turkish

model-releasesarxiv-cs-cl

arXiv:2512.23065v4 Announce Type: replace Abstract: The introduction of BERT established encoder-only transformer models as a foundational paradigm in natural language processing. Encoder-only models

The Ballon d'Or, legal division: On Harvey's Legal Agent Benchmark (LAB), Mistral Large 4 is the #1 oss model.

model-releasesmistral-ai--x

MistralAI announced that its model **Mistral Large 4** ranked as the #1 open‑source (OSS) model on **Harvey’s Legal Agent Benchmark (LAB)**, an evaluation benchmark for legal AI agents. The tweet post

Tree Navigation Without LLM Summaries: A Matched-Cost Study of Hierarchical Retrieval for Long-Document QA

arXiv:2610.06902v1 Announce Type: new Abstract: Retrieval-augmented generation grounds language models in external context, but for long documents flat top-k retrieval can cluster on a single region a

researcharxiv-cs-cl

Visual Abstention in Unified Multimodal Models

arXiv:2610.07887v1 Announce Type: cross Abstract: Unified multimodal models (UMMs) integrate understanding and generation, yet their generative behavior is rarely governed by what they understand abou

What Does Frechet Distance Measure? A Directional Decomposition

arXiv:2610.05518v1 Announce Type: cross Abstract: The Frechet distance is a de facto standard for evaluating generative models across domains, appearing as FID for images and FVD for videos. It summar

safetyarxiv-cs-lg

When Does AI Supervision Help? A Role-Aware Study of Network Fraud Decision Management with Blockchain Auditability

arXiv:2610.07434v1 Announce Type: new Abstract: When does a second artificial intelligence (AI) component improve a primary network-fraud decision rather than add operational burden? We study this que

A Bird's-Eye View of Iterative Reward Design

arXiv:2610.04364v1 Announce Type: new Abstract: Designing effective reward functions in RL typically requires substantial expertise and trial and error. Recent work automates this process with LLM-bas

Agentic-ZTA: A Multi-Agent Architecture for Autonomous Zero Trust Enforcement

arXiv:2610.05782v1 Announce Type: cross Abstract: Agentic AI is emerging as a promising paradigm for automating complex cybersecurity decisions, yet its use in enforcing zero trust introduces signific

safetyarxiv-cs-lg

AIProver: Agentic Auto-Formalization of Mathematical Research via Certificate-Driven Evolving Harness

arXiv:2610.05367v1 Announce Type: cross Abstract: Proof auto-formalization translates natural-language (NL) theorems and proofs into a formal language (FL) such as Lean, enabling mechanical verificati

All against the machine: the Solo score for rating skill in variable environments

arXiv:2610.04523v1 Announce Type: cross Abstract: We propose a distribution-free metric to rate individual skill in ``player-versus-environment'' settings, where participants face heterogeneous tasks

BabelFake: A Multilingual Audio-Visual DeepFake Benchmark

model-releasesarxiv-cs-cv

arXiv:2610.06339v1 Announce Type: new Abstract: Reliable and practical audio-visual DeepFake detection requires benchmarks that reflect diverse linguistic contexts and modern data synthesis pipelines

BARQ: Balanced Codebook Refinement for Low-Bit LLM Quantization

arXiv:2610.04490v1 Announce Type: new Abstract: As large language models (LLMs) grow in parameter count, model storage and parameter memory traffic have become major bottlenecks to efficient deploymen

Benchmarking CLIP for Zero-Shot Face and Periocular Gender Estimation

arXiv:2610.06102v1 Announce Type: new Abstract: We investigate CLIP for zero-shot gender estimation from full-face and periocular images. Three CLIP backbones are evaluated on 11,299 frontal images fr

safetyarxiv-cs-cv

Beyond In-Distribution Preservation: Recovering Generalization in Quantized VLAs via Vulnerability-Oriented Tuning

arXiv:2610.05745v1 Announce Type: new Abstract: Post-training quantization has been shown to preserve VLA performance under standard evaluation conditions, but whether it preserves the full-precision

safetyarxiv-cs-lg

Beyond the Good, the Bad, and the Ugly: Colormap Assessment through Data-Aware Perceptual Metric

arXiv:2610.03720v1 Announce Type: cross Abstract: Continuous colormaps are widely used to visualize scalar fields, and their quality is typically evaluated using measures such as discriminative power

CASE: Cost-Aware Stopping for Efficient Long-Video Agents

model-releasesarxiv-cs-cv

arXiv:2610.05400v1 Announce Type: cross Abstract: Long-video agents can actively gather question-relevant evidence, but they typically leave a central decision implicit: when has the agent seen enough

CHAOSMINING: Benchmarking Post-Hoc Attribution with Sparse Informative Features in High Dimensions

arXiv:2406.12150v2 Announce Type: replace Abstract: Post-hoc attribution is widely used to identify important model inputs, but evaluating whether these attributions identify truly informative feature

Agents

COMIC: Agentic Sketch Comedy Generation

arXiv:2603.11048v2 Announce Type: replace Abstract: We propose a fully automated AI system that produces short comedic videos similar to sketch shows such as Saturday Night Live. Starting from charact

agentsarxiv-cs-cv

Demonstration-Calibrated Port-Hamiltonian Retuning for Manipulation Policies

arXiv:2610.05755v1 Announce Type: new Abstract: Diffusion and VLA policies for manipulation are often deployed through downstream impedance controllers. The stiffness and damping gains of these contro

safetyarxiv-cs-ro

Dual Variational Autoencoders for Efficient Sim-to-Real Transfer in Low-Cost Robotic Navigation

arXiv:2610.06327v1 Announce Type: cross Abstract: Vision-based autonomous navigation for low-cost robots remains a fundamental challenge, primarily due to the significant gap between simulated trainin

safetyarxiv-cs-cv