Search: “benchmark”

Residual Drift Dominates Contradiction in Multi-Turn Constraint Reasoning

arXiv:2605.23940v1 Announce Type: new Abstract: How do multi-turn reasoning systems fail? The expected answer is logical contradiction, in which the system's maintained state becomes unsatisfiable. We

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Retrieved In-Context Principles from Previous Mistakes

DGX agent

arXiv:2407.05682v2 Announce Type: replace Abstract: In-context learning (ICL) has been instrumental in adapting Large Language Models (LLMs) to downstream tasks using correct input-output examples. Re

researcharxiv-cs-cl

26 May 2026

Revisiting Pre-Propagation GNNs: Robust Diffusion Operators and Hidden-State Re-Propagation

arXiv:2605.25111v1 Announce Type: new Abstract: Pre-propagation graph neural networks (PPGNNs) decouple node feature propagation from transformation: graph diffusion is performed once as preprocessing

Rewarding Structural Conformance of Reasoning using Process Mining

arXiv:2510.25065v3 Announce Type: replace Abstract: Recent advances in sparse reward policy gradient methods have enabled effective reinforcement learning (RL)-based language model post-training. Howe

Riemannian-Manifold Steering: Geometry-Aware Generative Autoencoders for Label-Free Steering

arXiv:2605.24942v1 Announce Type: cross Abstract: Steering a language model - intervening on its internal activations to change downstream behaviour - has recently expanded beyond linear interpolation

RoboManipBaselines: A Unified Framework for Imitation Learning in Robotic Manipulation across Real and Simulation Environments

model-releasesarxiv-cs-ro

arXiv:2509.17057v3 Announce Type: replace Abstract: We present RoboManipBaselines, an open-source software framework for imitation learning research in robotic manipulation. The framework supports the

SAE-FD: Sparse Autoencoder Feature Distillation for Continual Learning of Large Language Models

arXiv:2605.25525v1 Announce Type: new Abstract: Continual learning enables large language models to adapt to evolving tasks without retraining from scratch, yet catastrophic forgetting remains a centr

Schema-Grounded LLM Extraction for FHIR Patient Digital Twins

model-releasesarxiv-cs-cl

arXiv:2601.05847v2 Announce Type: replace Abstract: We revisit the problem of constructing interoperable patient digital twins from unstructured electronic health records (EHRs) and argue that the tas

Second Guess: Detecting Uncertainty Through Abstention and Answer Stability in Small Language Models

arXiv:2605.25394v1 Announce Type: new Abstract: Large language models often generate confident but incorrect answers rather than abstaining when uncertain. This problem is particularly acute for small

Selective Latent Thinking: Adaptive Compression of LLM Reasoning Chains

arXiv:2605.25745v1 Announce Type: new Abstract: Explicit chain-of-thought (CoT) reasoning substantially improves the reasoning ability of large language models (LLMs), but incurs high inference cost d

safetyarxiv-cs-cl

Self-Balancing Gradient Allocation for Heterogeneity-Aware Feature Generation in Click-Through Rate Prediction

arXiv:2605.24986v1 Announce Type: cross Abstract: Generative pre-training via discrete diffusion provides dense reconstruction supervision across all feature fields simultaneously, mitigating represen

SemanticZip: A Pilot Framework for Lossy Text Compression with LLMs as Semantic Decompressors

arXiv:2605.24541v1 Announce Type: cross Abstract: Text compression for large language model (LLM) systems is usually framed as token deletion, retrieval, summarization, or exact reconstruction. We stu

SentGraph: Hierarchical Sentence Graph for Multi-hop Retrieval-Augmented Question Answering

arXiv:2601.03014v3 Announce Type: replace-cross Abstract: Traditional Retrieval-Augmented Generation (RAG) effectively supports single-hop question answering with large language models but faces signi

Applications

Simulating Human Memory with Language Models

arXiv:2605.25680v1 Announce Type: cross Abstract: Language models are increasingly being deployed as user simulators, but their memory is far more reliable than that of real users. To measure this gap

applicationsarxiv-cs-ai

'Si'multaneous 'S'patial-'T'emporal Message Passing for Dynamic Graph Representation Learning

arXiv:2605.25548v1 Announce Type: cross Abstract: Dynamic graph neural networks (DGNNs) that operate on snapshot sequences typically fall into one of two categories. Temporal-first approaches build pe

SODE: Analyzing Social Dynamics in LLM Agents

arXiv:2605.23949v1 Announce Type: cross Abstract: As Large Language Models (LLMs) evolve into interactive agents, understanding their behavioral alignment within human social dynamics becomes essentia

SoK: A Comprehensive Security Analysis of Jailbreak Resilience in GPT and DeepSeek Models

arXiv:2506.18543v2 Announce Type: replace-cross Abstract: The rapid proliferation of Large Language Models (LLMs) has heightened concerns regarding their exposure to jailbreak attacks, which craft adv

SomaliBench Eval: Measuring English-to-Somali Refusal Gaps in Open-Weight Language Models

arXiv:2605.25420v1 Announce Type: cross Abstract: Large language model safety evaluation remains heavily English-centered, leaving low-resource languages under-measured even when models are deployed g

Spectral Retrieval: Multi-Scale Sinc Convolution over Token Embeddings for Localized Retrieval in LLM Multi-Agent Systems

arXiv:2605.24764v1 Announce Type: cross Abstract: [Abridged] - Spectral Retrieval is a plug-in re-ranking stage that interpolates between per-token MaxSim and mean-pool retrieval through a multi-scale

STaT: Resolving Shape Distortion in Non-Stationary Time Series via Tri-Modal Synergy

arXiv:2605.25943v1 Announce Type: new Abstract: Recent research in time series forecasting frequently investigates the integration of textual and visual modalities with numerical models to better navi

safetyarxiv-cs-lg

Stein Variational Ergodic Surface Coverage with SE(3) Constraints

model-releasesarxiv-cs-ro

arXiv:2603.09458v3 Announce Type: replace Abstract: Surface manipulation tasks require robots to generate trajectories that comprehensively cover complex 3D surfaces while maintaining precise end-effe

Stop Comparing LLM Agents Without Disclosing the Harness

arXiv:2605.23950v1 Announce Type: new Abstract: This position paper argues that, for long-horizon tasks evaluated across models with comparable frontier capability, the agent execution harness, namely

StructBreak: Structural Cognitive Overload-Induced Safety Failures in MLLMs

arXiv:2605.25534v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) excel at structural reasoning yet suffer from a sharp logical brittleness in structural consistency. We term th

Teaching large language models to reason like expert diagnosticians

arXiv:2509.12194v2 Announce Type: replace Abstract: Differential diagnosis is an iterative process that integrates patient information with broader medical knowledge. Clinical case series such as the

Test-Time Graph Search for Goal-Conditioned Reinforcement Learning

arXiv:2510.07257v2 Announce Type: replace Abstract: Offline goal-conditioned reinforcement learning (GCRL) often struggles with long-horizon tasks, where errors in value estimation accumulate and prod

Test-Time Self-Adaptive Conditioning for Stable Audio-Driven Talking-Head Generation

arXiv:2605.25488v1 Announce Type: cross Abstract: Audio-driven talking-head generation has achieved remarkable progress with recent models such as AniTalker, FLOAT, and Sonic. Despite their success, m

TGFormer: Towards Temporal Graph Transformer with Auto-Correlation Mechanism

arXiv:2605.24971v1 Announce Type: cross Abstract: The growing interest in Temporal Graph Neural Networks (TGNNs) stems from their ability to model complex dynamics and deliver superior performance. Ho

The Model Is Not the Product: A Dual-Pillar Architecture for Local-First Psychological Coaching

arXiv:2605.24411v1 Announce Type: new Abstract: Existing language model applications struggle to meet the demand for emotionally oriented support, primarily due to their inability to maintain deep, pe

The Perception-Physics Paradox: Probing Scientific Alignment with TC-Bench

arXiv:2605.24782v1 Announce Type: new Abstract: While Vision Foundation Models (VFMs) excel at predictive tasks on satellite imagery, their performance can arise from visual correlations rather than u

Agents

Tool-Call Dependency Structure is Linearly Decodable in LLM Agent Residual Streams

arXiv:2605.25310v1 Announce Type: new Abstract: Tool-using LLM agents produce trajectories whose calls form a directed dependency graph: earlier tool outputs supply arguments to later calls. Whether t

agentsarxiv-cs-cl

ToolRegistry: A Protocol-Agnostic Tool Management Library for Function-Calling LLMs

arXiv:2507.10593v3 Announce Type: replace-cross Abstract: Every LLM tool call is structurally an RPC -- a function name, JSON arguments, and a serialized result -- yet each protocol (native Python, MC

Topology-Driven Transferability Estimation of Medical Foundation Models for Segmentation

arXiv:2602.23916v2 Announce Type: replace-cross Abstract: The advent of large-scale self-supervised learning (SSL) has produced a vast zoo of medical foundation models. However, selecting optimal medi

Towards trustworthy agentic AI: a comprehensive survey of safety, robustness, privacy, and system security

arXiv:2605.23989v1 Announce Type: new Abstract: Agentic AI systems -- Large Language Models (LLMs) augmented with planning, tool use, memory, and long-horizon interactions -- can execute complex tasks

TRAFA: Anticipating User Actions to Reduce Errors in Procedural Tasks with Predictive Feedback

arXiv:2605.24526v1 Announce Type: cross Abstract: Interactive assistance systems typically provide feedback after an action has been completed, supporting error recovery but not preventing the error i

Trajectory-Based Difficulty Scoring for Reliable Learning on Tabular Data

arXiv:2605.24680v1 Announce Type: new Abstract: Gradient-boosted trees achieve strong performance on tabular data, yet often leave a long tail of poorly predicted instances. We introduce a Trajectory-

Trust-Aware Joint Feature-Prediction Discrepancy for Robust Domain Adaptation

arXiv:2605.25119v1 Announce Type: cross Abstract: Domain adaptation aims to mitigate performance degradation caused by distribution shifts between a labeled source domain and an unlabeled or sparsely

Truthful Online Preference Aggregation for LLM Fine-Tuning in Mobile Crowdsourcing

arXiv:2605.24052v1 Announce Type: cross Abstract: To better serve users' demands in mobile applications (e.g., navigation), mobile crowdsourcing platforms can iteratively align large language model (L

Uncertainty Decomposition via Cyclical SG-MCMC and Soft-label Learning for Subjective NLP

arXiv:2605.24773v1 Announce Type: new Abstract: Annotator disagreement in emotion classification reflects ambiguity intrinsic to emotion concepts and is essential for predictor-quality assessment in s

Tutorials

Uncertainty Reasoning with Large Language Models for Explainable Disease Diagnosis

arXiv:2605.25566v1 Announce Type: new Abstract: Clinical decision-making requires reasoning over incomplete, imprecise, and linguistically expressed patient narratives. While large language models (LL

tutorialsarxiv-cs-ai

Uncovering Autoregressive LLM Knowledge of Thematic Fit in Event Representation

arXiv:2410.15173v4 Announce Type: replace-cross Abstract: The thematic fit estimation task measures semantic arguments' compatibility with a given semantic role for a given predicate. We investigate i

Understanding and Mitigating Premature Confidence for Better LLM Reasoning

arXiv:2605.24396v1 Announce Type: new Abstract: Long chains of thought (CoT) from current language models frequently contain logical gaps and unjustified leaps, limiting the gains from additional test

V3H: View Variation and View Heredity for Incomplete Multi-view Clustering

arXiv:2011.11194v4 Announce Type: replace Abstract: Real data often appear in the form of multiple incomplete views. Incomplete multi-view clustering is an effective method to integrate these incomple

VaaWIT: Visual-Aware Adaptation of Large Language Models for Multilingual Web Image Translation

arXiv:2605.24675v1 Announce Type: cross Abstract: Translating text embedded in Web images is crucial for improving content accessibility and cross-lingual information retrieval, particularly within so

Applications

VectorArk: Learning Practical Image Vectorization with Rounded Polygon Representation

arXiv:2605.24398v1 Announce Type: cross Abstract: Recent vision-language model (VLM)-based approaches have achieved impressive results on image vectorization tasks. However, they are typically evaluat

applicationsarxiv-cs-ai

Visual-Redundancy-Controlled Parallel Decoding for Diffusion-Based Multimodal Large Language Models

arXiv:2605.25820v1 Announce Type: new Abstract: Diffusion-based multimodal large language models (dMLLMs) decode by iteratively predicting tokens at multiple masked positions in parallel. This turns e

What Makes a Medical Checker Trainable? Diagnosing Signal Collapse and Reward Hacking in Checker-Guided RAG for Biomedical QA

model-releasesarxiv-cs-cl

arXiv:2605.25988v1 Announce Type: new Abstract: Medical RAG needs evidence-grounded claims, so plugging a claim-level NLI checker into retrieval-augmented RL is intuitive. extbf{We find that the check

What Questions Should Robots Be Able to Answer? A Dataset of User Questions for Explainable Robotics

arXiv:2510.16435v2 Announce Type: replace-cross Abstract: With the growing use of large language models and conversational interfaces in human-robot interaction, robots' ability to answer user questio

researcharxiv-cs-cl

When Can We Trust Early Warnings? Leakage-Excluded Early Outcome Prediction from LMS Interaction Logs

arXiv:2605.25794v1 Announce Type: new Abstract: Early-warning models built from Learning Management System (LMS) logs aim to predict end-of-course outcomes early enough to enable timely learner suppor