A training-free post-hoc code watermark embeds 24-bit identifiers via multi-channel spread-spectrum encoding over naming conventions and semantic pattern pairs, with majority voting and Reed-Solomon recovery.
hub Canonical reference
Batch-ICL: Effective, efficient, and order-agnostic in-context learning
Canonical reference. 100% of citing Pith papers cite this work as background.
28
Pith papers citing it
Background
100% of classified citations
hub tools
citation-role summary
background 7
citation-polarity summary
roles
background 7polarities
background 7representative citing papers
MLLMs drop from over 85% accuracy on action presence to under 50% on matched action-denial videos, exposing a causal verification gap that causal graph prompts partially close.
White-box method ReXTrust achieves highest AUC (peak 93.0) on Gut-VLM across five VLMs, outperforming alternatives by statistically significant margins while black-box and some gray-box methods collapse on certain models.
A test-time zeroth-order optimization of prompt embeddings using a bounded self-supervised proxy from demonstration log-probabilities improves ICL accuracy and correlates with gains across tasks.
DiffCodeGen clusters code candidates by behavioral similarity from fuzzing-synthesized inputs and selects the largest cluster's medoid, matching or exceeding prior test-time scaling methods with far less token and time cost.
Hydra enables asynchronous static error checking and targeted checkpoint-rollback repair during LLM code generation, cutting latency by up to 71% and token use by up to 70% versus post-hoc repair on C/C++ tasks.
SimDiff uses similarity and difference metrics to prune LLM layers more effectively than cosine similarity alone, retaining over 91% performance at 25% pruning on LLaMA2-7B.
LLM-based dense retrievers generalize better when instruction-tuned but pay a specialization tax when optimized for reasoning; they resist typos and corpus poisoning better than encoder-only baselines yet remain vulnerable to semantic perturbations, with larger models and certain embedding geometry,
CurateEvo evolves executable data-curation code using failed agent trajectories, improving post-training performance by 3.2 and 2.7 points over baselines on labeled and wild data respectively.
Introduces a benchmark for MLLM-based chart data extraction from unlabeled images and a human-centered training framework that reaches SOTA numerical accuracy with a 7B model.
ActPlane introduces an OS-kernel policy engine using an information-flow control DSL and eBPF to enforce agent harness policies, achieving better compliance on indirect paths with 1.9-8.4% overhead.
PRISM benchmark finds LLMs match or exceed humans on isolated review dimensions like novelty verification but none achieve the balanced performance of human reviewers across depth, flaw prioritization, and constructiveness.
PromptAudit evaluates five prompting strategies across five LLMs on 1000 CVEs and finds chain-of-thought prompting yields the strongest overall performance while adaptive chain-of-thought and self-consistency reduce effective results.
An Explicit Logic Channel of LLM, VFM and probabilistic inference validates and improves zero-shot MLLMs via Consistency Rate without ground-truth labels.
AI agents can generate code in a capability-safe Scala dialect that statically prevents information leakage and malicious side effects while preserving task performance.
VeriOS-Agent is an OS agent that proactively queries humans in untrustworthy scenarios via a query-driven framework and three-stage training, achieving 19.72% higher step-wise success rate over baselines while preserving normal performance.
GPT produces click distributions significantly different from real humans in 53% of UX first-click tasks, with prompting techniques like personas and chain-of-thought failing to improve alignment.
MHSA mitigates hallucinations in LVLMs by training an MLP to steer cross-modal attention, extending detection work to mitigation via attention replacement at inference.
Passages made from high-convergence sentences improve LLM performance on inferential questions compared to cosine similarity selection.
CUCI-Net abstracts context-utterance dependency into an interpretation cue that combines local modality signals with global context and feeds it into the final multimodal interaction for context-conditioned predictions.
STAR is a semantic-tuned and tail-adaptive retriever for GraphRAG that uses cross-attention interaction learning and path-weighted contrastive learning to mitigate Semantic Shortcut Bias and Long-Tail Path Bias, reporting 1.8% retrieval and 2.2% QA gains.
A multi-agent generate-validate-revise framework reduces failures in realism and authenticity for LLM-personalized math problems, with one iteration helping and different strategies varying by criterion.
The paper unifies perspectives on Long CoT in reasoning LLMs by introducing a taxonomy, detailing characteristics of deep reasoning and reflection, and discussing emergence phenomena and future directions.
A survey of 87 agents for computer use and 33 datasets that introduces a three-dimensional taxonomy across domain, interaction, and agent perspectives and identifies six research gaps.
citing papers explorer
-
Multi-Channel Spread-Spectrum Code Watermarking
A training-free post-hoc code watermark embeds 24-bit identifiers via multi-channel spread-spectrum encoding over naming conventions and semantic pattern pairs, with majority voting and Reed-Solomon recovery.
-
Learning to Deny: Action Denial in Multimodal Large Language Models
MLLMs drop from over 85% accuracy on action presence to under 50% on matched action-denial videos, exposing a causal verification gap that causal graph prompts partially close.
-
A Benchmark for Hallucination Detection in VLMs for Gastrointestinal Endoscopy
White-box method ReXTrust achieves highest AUC (peak 93.0) on Gut-VLM across five VLMs, outperforming alternatives by statistically significant margins while black-box and some gray-box methods collapse on certain models.
-
Self-Improving In-Context Learning
A test-time zeroth-order optimization of prompt embeddings using a bounded self-supervised proxy from demonstration log-probabilities improves ICL accuracy and correlates with gains across tasks.
-
Code Generation by Differential Test Time Scaling
DiffCodeGen clusters code candidates by behavioral similarity from fuzzing-synthesized inputs and selects the largest cluster's medoid, matching or exceeding prior test-time scaling methods with far less token and time cost.
-
Hydra: Efficient, Correct Code Generation via Checkpoint-and-Rollback Support
Hydra enables asynchronous static error checking and targeted checkpoint-rollback repair during LLM code generation, cutting latency by up to 71% and token use by up to 70% versus post-hoc repair on C/C++ tasks.
-
SimDiff: Depth Pruning via Similarity and Difference
SimDiff uses similarity and difference metrics to prune LLM layers more effectively than cosine similarity alone, retaining over 91% performance at 25% pruning on LLaMA2-7B.
-
On the Robustness of LLM-Based Dense Retrievers: A Systematic Analysis of Generalizability and Stability
LLM-based dense retrievers generalize better when instruction-tuned but pay a specialization tax when optimized for reasoning; they resist typos and corpus poisoning better than encoder-only baselines yet remain vulnerable to semantic perturbations, with larger models and certain embedding geometry,
-
CurateEvo: Data-Curation Evolving for Agentic Post-Training
CurateEvo evolves executable data-curation code using failed agent trajectories, improving post-training performance by 3.2 and 2.7 points over baselines on labeled and wild data respectively.
-
Making Multimodal LLMs Reliable Chart Data Extractors: A Benchmark and Training Framework
Introduces a benchmark for MLLM-based chart data extraction from unlabeled images and a human-centered training framework that reaches SOTA numerical accuracy with a 7B model.
-
ActPlane: Programmable OS-Level Policy Enforcement for Agent Harnesses
ActPlane introduces an OS-kernel policy engine using an information-flow control DSL and eBPF to enforce agent harness policies, achieving better compliance on indirect paths with 1.9-8.4% overhead.
-
PRISM: A Multi-Dimensional Benchmark for Evaluating LLM Peer Reviewers
PRISM benchmark finds LLMs match or exceed humans on isolated review dimensions like novelty verification but none achieve the balanced performance of human reviewers across depth, flaw prioritization, and constructiveness.
-
PromptAudit: Auditing Prompt Sensitivity in LLM-Based Vulnerability Detection
PromptAudit evaluates five prompting strategies across five LLMs on 1000 CVEs and finds chain-of-thought prompting yields the strongest overall performance while adaptive chain-of-thought and self-consistency reduce effective results.
-
Explicit Logic Channel for Validation and Enhancement of MLLMs on Zero-Shot Tasks
An Explicit Logic Channel of LLM, VFM and probabilistic inference validates and improves zero-shot MLLMs via Consistency Rate without ground-truth labels.
-
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.
-
VeriOS: Query-Driven Proactive Human-Agent-GUI Interaction for Trustworthy OS Agents
VeriOS-Agent is an OS agent that proactively queries humans in untrustworthy scenarios via a query-driven framework and three-stage training, achieving 19.72% higher step-wise success rate over baselines while preserving normal performance.
-
What Would GPT Click: Practical Effects of Human-AI Behavioral Misalignment and the Cost of Synthetic Participants in User Experience
GPT produces click distributions significantly different from real humans in 53% of UX first-click tasks, with prompting techniques like personas and chain-of-thought failing to improve alignment.
-
MHSA: A Lightweight Framework for Mitigating Hallucinations via Steered Attention in LVLMs
MHSA mitigates hallucinations in LVLMs by training an MLP to steer cross-modal attention, extending detection work to mitigation via attention replacement at inference.
-
Context Convergence Improves Answering Inferential Questions
Passages made from high-convergence sentences improve LLM performance on inferential questions compared to cosine similarity selection.
-
Beyond Isolated Utterances: Cue-Guided Interaction for Context-Dependent Conversational Multimodal Understanding
CUCI-Net abstracts context-utterance dependency into an interpretation cue that combines local modality signals with global context and feeds it into the final multimodal interaction for context-conditioned predictions.
-
STAR: Semantic-Tuned and Tail-Adaptive Retriever for Graph-Augmented Generation
STAR is a semantic-tuned and tail-adaptive retriever for GraphRAG that uses cross-attention interaction learning and path-weighted contrastive learning to mitigate Semantic Shortcut Bias and Long-Tail Path Bias, reporting 1.8% retrieval and 2.2% QA gains.
-
A Multi-Agent Approach to Validate and Refine LLM-Generated Personalized Math Problems
A multi-agent generate-validate-revise framework reduces failures in realism and authenticity for LLM-personalized math problems, with one iteration helping and different strategies varying by criterion.
-
Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models
The paper unifies perspectives on Long CoT in reasoning LLMs by introducing a taxonomy, detailing characteristics of deep reasoning and reflection, and discussing emergence phenomena and future directions.
-
A Comprehensive Survey of Agents for Computer Use: Foundations, Challenges, and Future Directions
A survey of 87 agents for computer use and 33 datasets that introduces a three-dimensional taxonomy across domain, interaction, and agent perspectives and identifies six research gaps.
-
Reasoning4Sciences: Bridging Reasoning Language Models to All Scientific Branches
A survey of reasoning language model adoption across 28 ERC scientific disciplines finds large maturity gaps, especially when only public resources are counted.
-
RAG-Enhanced Large Language Models for Dynamic Content Expiration Prediction in Web Search
An LLM framework with RAG predicts query-specific validity horizons for web content expiration and shows gains in production A/B tests.
-
Multilingual Vision-Language Models, A Survey
The survey identifies a key tension in multilingual vision-language models between language neutrality via contrastive learning and cultural awareness via diverse data, with most benchmarks relying on translation-based evaluation.
- IDRBench: Understanding the Capability of Large Language Models on Interdisciplinary Research