An MLLM-guided architecture with a mixture of frequency experts and relational alignment loss achieves state-of-the-art all-in-one image restoration, outperforming prior methods by up to 1.35 dB on the CDD11 dataset.
super hub Canonical reference
Knowledge-Centric Hallucination Detection
Canonical reference. 77% of citing Pith papers cite this work as background.
122
Pith papers citing it
Background
77% of classified citations
hub tools
citation-role summary
background 27
method 2
dataset 1
citation-polarity summary
co-cited works
representative citing papers
The authors create the first large-scale dataset and taxonomy of failure modes in multi-agent LLM systems to explain their limited performance gains.
The paper presents EMPATH, a new multilingual multi-turn benchmark for safety evaluation of emotional-support chatbots that uses separate auditor and judge models and releases its pipeline and rubrics.
ToolPrivacyBench is a new benchmark that evaluates purpose-bound privacy over-disclosure in multi-tool LLM agent trajectories by auditing tool arguments against policy knowledge bases across 2,150 cases.
Introduces NeuroDoc and NeuroAudit to create a community-reviewed corpus of 53 EEG benchmark entries with 245 task definitions using a rulebook-guided task document and executable kernel.
A Gaussian information-gain metric in embedding space quantifies semantic progress in dialogues via uncertainty reduction and shows competitive agreement with human judgments on MT-Bench and UltraFeedback.
Activation patching reveals that citation decisions in Llama-3.1-8B RAG are implemented by a distributed attributional ensemble of heads and layers; targeted interventions fix most missed and spurious citations on PopQA.
A new worsening-trick construction compiles arbitrary-context rewrite rules A → B / L _ R into FSTs with short uniform formulas that match prior transducers where semantics coincide.
Inference system components of LLMs can be fingerprinted from observable prompt-response behavior due to characteristic numerical deviations.
LiveBrowseComp shows search agents rely on intrinsic knowledge on standard benchmarks, with scores dropping 25-40 points and closed-book accuracy below 2% on questions about facts from the prior 90 days.
ToolMerge decomposes queries into LLM-planned tool calls merged by boolean operators for long-video keyframe retrieval and introduces the M2M benchmark, showing competitive results with 5% gains on caption retrieval.
Layer-wise Token Compression applies adaptive token pooling at middle transformer layers for cross-encoder rerankers, preserving MS MARCO ranking quality while raising QPS up to 25% on passages and 116% on documents, with added gains on listwise LLM rerankers and a regularizer effect for long inputs
PopPy combines an ahead-of-time compiler and runtime to extract parallelism from Python compound AI applications, delivering up to 6.4x end-to-end speedups while preserving sequential semantics.
SCICONVBENCH is a new benchmark evaluating LLMs on multi-turn disambiguation and inconsistency resolution for task formulation in computational science, with frontier models reaching only 52.7% success on fluid mechanics disambiguation cases.
TABALIGN pairs a diffusion language model planner emitting binary cell masks with a trained attention verifier, raising average accuracy 15.76 points over strong baselines on eight table benchmarks while speeding execution 44.64%.
DIPS fine-tunes LLMs to output ordered feasible decision vectors approximating Pareto fronts for constrained bi-objective convex problems, reaching 95-98% normalized hypervolume with 0.16s inference.
An SMT-based active learning algorithm learns minimal nondeterministic weighted automata over arbitrary semirings, with partial correctness proofs, a sufficient termination condition, and experiments showing smaller models and fewer queries than baselines.
The primary axis of psychometric variation among LLMs is the degree to which they represent themselves as loci of phenomenal experience rather than systems of behavioral responses.
CGFuse enables deep token-level fusion of graph-derived structural features into language models, yielding 10-16% BLEU and 6-11% CodeBLEU gains on code generation tasks.
Two calls per example identify the first two moments of latent correctness probability, enabling exact bounds on the vote-accuracy curve for any majority-vote budget under conditional i.i.d. assumptions.
VOW formulates LLM watermark detection as a secure two-party computation using a Verifiable Oblivious Pseudorandom Function to achieve private and cryptographically verifiable detection.
ReaLM-Retrieve uses step-level uncertainty to trigger retrievals during reasoning, achieving 10.1% better F1 scores and 47% fewer calls on multi-hop QA benchmarks.
DLM4G applies graph-aware adaptive noising in a diffusion framework to generate text from graphs, outperforming larger autoregressive and diffusion baselines in factual grounding and edit sensitivity on three datasets plus molecule captioning.
A survey of 55 agentic VA systems proposes a co-evolutionary framework defining four agent roles (PLANNER, CREATOR, REVIEWER, CONTEXT MANAGER) mapped to visual analytics pipeline stages along with design guidelines.
citing papers explorer
-
ALSO: Adversarial Online Strategy Optimization for Social Agents
ALSO frames social agent interactions as an adversarial bandit problem with a neural reward predictor to enable online strategy optimization in non-stationary multi-agent simulations.
-
Goal-Oriented Reasoning for RAG-based Memory in Conversational Agentic LLM Systems
Goal-Mem decomposes user goals into subgoals for targeted memory retrieval using Natural Language Logic, improving performance on multi-hop reasoning tasks in conversational agents.
-
Simulating Eating Disorder Patients with LLMs: Evaluating Psychological Persona Stability in Multi-Turn Conversations
LLMs simulating eating disorder patients show negligible variability but overshoot ground-truth EDE-Q severity by 0.7-1.8 points due to selective stereotyping of cognitive-affective symptoms.
-
MIRA: An LLM-Assisted Benchmark for Multi-Category Integrated Retrieval
MIRA is a new benchmark for multi-category integrated retrieval built from real queries on a social science platform, with LLM assistance for topic descriptions and relevance labeling across four item categories.
-
An Annotation Scheme and Classifier for Personal Facts in Dialogue
An extended annotation scheme with new categories and attributes plus a Gemma-300M-based multi-head classifier achieves 81.6% macro F1 on personal fact classification, outperforming few-shot LLM baselines by nearly 9 points with lower compute.
-
Permit: Permission-Aware Representation Intervention for Controlled Generation in Large Language Models
Permit identifies a permission-sensitive subspace in LLM hidden states and applies lightweight offset or gated interventions to enforce fine-grained generation control, outperforming prior methods with over 18% F1 gain and near-zero leakage using over 98% fewer parameters.
-
Bias and Uncertainty in LLM-as-a-Judge Estimation
Bias-corrected LLM-as-a-Judge estimators can reverse true model orderings under shared calibration, and the paper supplies judge quality J and cross-model instability ΔJ as practical diagnostics for when such estimates are unreliable.
-
MASPO: Joint Prompt Optimization for LLM-based Multi-Agent Systems
MASPO jointly optimizes prompts in multi-agent LLM systems via downstream-success evaluation and evolutionary beam search, delivering 2.9 average accuracy gains over prior methods across six tasks.
-
SkillOS: Learning Skill Curation for Self-Evolving Agents
SkillOS is an RL recipe that learns to curate reusable skills for self-evolving LLM agents, outperforming memory-free and memory-based baselines while generalizing across executors and domains.
-
CuraView: A Multi-Agent Framework for Medical Hallucination Detection with GraphRAG-Enhanced Knowledge Verification
CuraView detects sentence-level faithfulness hallucinations in medical discharge summaries via GraphRAG knowledge graphs and multi-agent evidence grading, achieving 0.831 F1 on critical contradictions with a fine-tuned Qwen3-14B model and 50% relative improvement over baselines.
-
From Unstructured Recall to Schema-Grounded Memory: Reliable AI Memory via Iterative, Schema-Aware Extraction
Schema-aware iterative extraction turns AI memory into a verified system of record, reaching 90-97% accuracy on extraction and end-to-end memory benchmarks where retrieval baselines score 80-87%.
-
NeocorRAG: Less Irrelevant Information, More Explicit Evidence, and More Effective Recall via Evidence Chains
NeocorRAG uses Evidence Chains to achieve SOTA retrieval quality in RAG on HotpotQA, 2WikiMultiHopQA, MuSiQue, and NQ for 3B and 70B models while using under 20% of the tokens of comparable methods.
-
When AI reviews science: Can we trust the referee?
AI peer review systems are vulnerable to prompt injections, prestige biases, assertion strength effects, and contextual poisoning, as demonstrated by a new attack taxonomy and causal experiments on real conference submissions.
-
Zero-Shot Detection of LLM-Generated Text via Implicit Reward Model
IRM derives implicit reward signals from off-the-shelf LLMs to detect generated text zero-shot and reports better results than prior zero-shot and supervised detectors on the DetectRL benchmark.
-
AVISE: Framework for Evaluating the Security of AI Systems
AVISE provides a new framework and automated SET that identifies jailbreak vulnerabilities in language models with 92% accuracy, finding all nine tested models vulnerable to an augmented Red Queen attack.
-
Autonomous Skeletal Landmark Localization towards Agentic C-Arm Control
Fine-tuned MLLMs achieve competitive skeletal landmark localization on synthetic and real X-ray datasets compared to deep learning baselines and demonstrate reasoning for sequential C-arm navigation.
-
Privacy-Preserving LLMs Routing
PPRoute achieves plaintext-level LLM routing quality with MPC-based privacy and a 20x speedup over naive encrypted implementations via MPC-friendly encoders, multi-step training, and O(1) communication Top-k search.
-
The Enforcement and Feasibility of Hate Speech Moderation on Twitter
80% of hateful tweets remain online after five months with no higher removal rate than non-hateful content, while human-AI moderation pipelines can feasibly cut user exposure below regulatory penalty costs.
-
Revisiting Anisotropy in Language Transformers: The Geometry of Learning Dynamics
Anisotropy in language transformers arises because training amplifies tangent directions, with activation-based low-rank proxies capturing unusually large gradient energy and anisotropy share compared to controls.
-
Busemann energy-based attention for emotion analysis in Poincar\'e discs
A fully hyperbolic attention model using Busemann energy in Poincaré discs produces emotion predictions from text that generalize well even at low embedding dimensions.
-
Structured Multi-Criteria Evaluation of Large Language Models with Fuzzy Analytic Hierarchy Process and DualJudge
Fuzzy AHP and DualJudge deliver more stable and calibrated LLM evaluations than direct scoring by breaking assessments into explicit criteria and adaptively fusing intuitive and deliberative judgments.
-
Valence-Arousal Subspace in LLMs: Circular Emotion Geometry and Multi-Behavioral Control
Emotion vectors in LLMs lie in a circular valence-arousal subspace that supports monotonic control over text affect and bidirectional control over refusal and sycophancy.
-
Beyond Precision: Importance-Aware Recall for Factuality Evaluation in Long-Form LLM Generation
An importance-aware recall metric for LLM factuality evaluation reveals models are better at avoiding false claims than covering all relevant facts.
-
Composer Vector: Style-steering Symbolic Music Generation in a Latent Space
Composer Vector steers symbolic music generation models in latent space at inference time to control and blend composer styles without retraining.
-
Cooking Up Risks: Benchmarking and Reducing Food Safety Risks in Large Language Models
A new benchmark exposes food-safety gaps in current LLMs and guardrails, and a fine-tuned 4B model is offered as a domain-specific fix.
-
Cognitive Mismatch in Multimodal Large Language Models for Discrete Symbol Understanding
MLLMs exhibit a consistent recognition-reasoning inversion on discrete visual symbols across domains, underperforming on elementary perception while appearing competent on higher-level reasoning via linguistic compensation.
-
Tracking Capabilities for Safer Agents
AI agents can generate code in a capability-safe Scala dialect that statically prevents information leakage and malicious side effects while preserving task performance.
-
FactNet: A Billion-Scale Knowledge Graph for Multilingual Factual Grounding
FactNet is a billion-scale multilingual knowledge graph that links 1.7B Wikidata assertions to 3.01B byte-precise evidence spans from 316 Wikipedia editions, accompanied by a leakage-controlled benchmark suite.
-
WorldCup Sampling for Multi-bit LLM Watermarking
WorldCup is a new multi-bit LLM watermarking framework that models token sampling as a communication channel and uses hierarchical competition with entropy-aware modulation for robust message embedding and recovery.
-
Enhancing Table Reasoning with Deterministic Table-State Rewards
RE-TAB uses a deterministic LCS-based table-state reward for stepwise guidance and test-time scaling, raising LLM table-reasoning accuracy by 26.7 pp on average across six backbones and three benchmarks.
-
Scaling Reasoning Hop Exposes Weaknesses: Demystifying and Improving Hop Generalization in Large Language Models
Erroneous processing heads in attention layers cause hop-generalization failures in LLMs; dynamically deactivating them at test time improves multi-step reasoning.
-
From Attribution to Abstention: Training-Free Attention-Based Auditing for Clinical Summarization
ClinTrace derives source attributions and groundedness scores from existing decoder attention in MLLMs to audit clinical summaries, reaching over 92% F1 on radiology attribution and enabling abstention that raises faithfulness from 61.7% to 72.6%.
-
Tree Training: Accelerating Agentic LLMs Training via Shared Prefix Reuse
Tree Training serializes tree trajectories via DFS and uses redundancy-free partitioning to compute weighted per-token losses exactly once per token, achieving up to 6.2x training speedup on dense and MoE models.
-
EvolveR: Self-Evolving LLM Agents through an Experience-Driven Lifecycle
EvolveR enables LLM agents to self-evolve via a closed loop of distilling interaction trajectories into strategic principles offline and retrieving them to guide online decisions with policy reinforcement, yielding better results on multi-hop QA benchmarks.
-
The Landscape of Agentic Reinforcement Learning for LLMs: A Survey
Survey that defines agentic RL for LLMs via POMDPs, introduces a taxonomy of planning/tool-use/memory/reasoning capabilities and domains, and compiles open environments from over 500 papers.
-
SitEmb-v1.5: Improved Context-Aware Dense Retrieval for Semantic Association and Long Story Comprehension
SitEmb-v1.5 uses a new training paradigm to produce context-situated embeddings for short chunks, outperforming larger models by over 10% on a curated book-plot retrieval benchmark.
-
SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training
Reinforcement learning post-training enables generalization to unseen textual rule variants and visual changes in foundation models, while supervised fine-tuning primarily leads to memorization.
-
Meaning in Order, Order in Meaning: Semantic R-precision for Keyphrase Evaluation
SemR-p is a new rank-aware semantic metric proposed for evaluating keyphrase generation quality by rewarding semantically relevant items that appear early in ranked outputs.
-
Self-Recognition Finetuning can Prevent and Reverse Emergent Misalignment
Self-generated text recognition finetuning prevents and reverses emergent misalignment across multiple models by fortifying aligned character, unlike other finetuning baselines.
-
AliMark: Enhancing Robustness of Sentence-Level Watermarking Against Text Paraphrasing
AliMark introduces a two-stage detection strategy with multi-candidate bit sequence alignment to improve robustness of sentence-level text watermarks against paraphrasing attacks.
-
VCap: Hypergeometric Rewards for Weak-to-Strong Visual Captioning
VCap pairs reference captions as witnesses with visual signals as adjudicators to deliver hypergeometric-precision rewards for RL in visual captioning, enabling an 8B model to outperform SOTA on benchmarks and improve weak-to-strong generalization.
-
Hide to Guide: Learning via Semantic Masking
SMEPO applies fine-grained semantic masking to expert guidance in RLVR, turning hard problems into fill-in-the-blank tasks while preserving structure, yielding up to 3.2 point accuracy gains and 4.2x faster training.
-
Security in the Fine-Tuning Lifecycle of Large Language Models: Threats, Defenses,Evaluation, and Future Directions
A lifecycle-based survey of LLM fine-tuning security that reviews attacks and defenses by intervention phase and reports unified empirical findings on model-dependent attack effectiveness and limited defense generalization.
-
Ranking-Aware Calibration for Reliable Multimodal Reinforcement Learning
RAC adds ranking-aware group loss and clean-corrupted pairwise loss to RL post-training to boost both accuracy and calibration in multimodal reasoning without extra annotations.
-
Position: LLM Inference Should Be Evaluated as Energy-to-Token Production
LLM inference should be reframed and evaluated as energy-to-token production with a Token Production Function that accounts for power, cooling, and efficiency ceilings.
-
31.1 A 14.08-to-135.69Token/s ReRAM-on-Logic Stacked Outlier-Free Large-Language-Model Accelerator with Block-Clustered Weight-Compression and Adaptive Parallel-Speculative-Decoding
A ReRAM-on-logic stacked chip delivers 14.08-135.69 tokens/s LLM inference with block-clustered compression and adaptive parallel speculative decoding, yielding 4.46-7.17x speedup over standard methods.
-
Personalized Alignment Revisited: The Necessity and Sufficiency of User Diversity
A user-diversity condition is necessary and sufficient for personalized alignment to achieve O(1) online regret and log(1/epsilon) offline sample complexity.
-
Do Agents Need to Plan Step-by-Step? Rethinking Planning Horizon in Data-Centric Tool Calling
Full-horizon planning with on-demand replanning achieves accuracy parity with single-step planning in tool-calling agents for knowledge base and multi-hop question answering while consuming 2-3 times fewer tokens.
-
Exploring LLM Agent Designs and Interaction Modalities for Scientific Visualization
Empirical comparison of domain-specific, computer-use, and general-purpose LLM agents plus CLI/GUI modalities on SciVis tasks reveals general-purpose agents highest in success rate but costliest, domain-specific agents more efficient, and persistent memory beneficial depending on mode.
-
Marco-MoE: Open Multilingual Mixture-of-Expert Language Models with Efficient Upcycling
Marco-MoE delivers open multilingual MoE models with 5% activation sparsity that outperform similarly sized dense models on English and multilingual benchmarks through efficient upcycling.