TILens Daily Edition
2026-09-28
Filters: topic=ai-research, github=hidden

Stats: 168 articles, 6 sources, 161 research papers
Top topics: AI Research, AI, Information Retrieval

## News

- Gates’s “Billion Deaths” Warning Is Really a Fight Over Who Regulates AI
  Source: the-neuron-ai
  Topic: AI (+AI Research)
  URL: https://www.theneuron.ai/news/gates-billion-deaths-warning-who-regulates-ai/

- Muse Is a Hit. Now Meta Wants an Enterprise AI Business
  Source: the-neuron-ai
  Topic: AI (+AI Research)
  URL: https://www.theneuron.ai/news/muse-is-a-hit-now-meta-wants-an-enterprise-ai-business/

- NVIDIA’s Agent Safety Bet Goes Beyond Model Guardrails
  Source: the-neuron-ai
  Topic: AI (+AI Research)
  URL: https://www.theneuron.ai/news/nvidia-agent-safety-bet-beyond-model-guardrails/

- Claude Spotted Something Scientists Missed. Now Comes the Reproducibility Test
  Source: the-neuron-ai
  Topic: AI (+AI Research)
  URL: https://www.theneuron.ai/news/claude-spotted-something-scientists-missed-now-comes-reproducibility-test/

- 😸 Did OpenAI lose control? 🚨
  Source: the-neuron-ai
  Topic: AI (+AI Research)
  URL: https://www.theneuron.ai/newsletter/did-openai-lose-control/

- Can Errors From an AI Summary Distort Memory? A New Study Says Yes
  Source: the-neuron-ai
  Topic: AI (+AI Research)
  URL: https://www.theneuron.ai/news/ai-summary-errors-distort-memory-study/

- AI Weekly Issue #533: Meta tested human callers behind its AI phone agent
  Source: ai-weekly
  Topic: AI (+AI Research)
  URL: https://aiweekly.co/issues/meta-tested-human-callers-behind-its-ai-phone-agent

## Research

- Guided Merge Sort : An Optimized Sorting that Picks the Best from Ordinary and Multi-Way Merge Sort Algorithms
  Source: towards-data-science
  Topic: AI Research (+AI, Python)
  URL: https://towardsdatascience.com/guided-merge-sort-an-optimized-sorting-that-picks-the-best-from-ordinary-and-multi-way-merge-sort-algorithms/

- How to Make Your Own JEV Model from an Open LLM
  Source: towards-data-science
  Topic: AI Research (+AI, Python)
  URL: https://towardsdatascience.com/how-to-make-your-own-jev-model-from-an-open-llm/

- The AI That Learned to Understand Long After It Stopped Trying
  Source: towards-data-science
  Topic: AI Research (+AI, Python)
  URL: https://towardsdatascience.com/the-ai-that-learned-to-understand-long-after-it-stopped-trying/

- How to Catch Data Drift When Every Feature Looks Normal
  Source: towards-data-science
  Topic: AI Research (+AI, Python)
  URL: https://towardsdatascience.com/how-to-catch-data-drift-when-every-feature-looks-normal/

- MM-ContextFold: Context Folding for Multimodal Agentic Retrieval
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research, JavaScript, React)
  URL: https://arxiv.org/abs/2609.23121

- AutoResearch at Production Scale: Failure Modes and a Multi-Agent Framework
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30541

- AgentRecommender: LLM Agents Enable Customizable Recommender Systems on the User Side
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.31166

- FlyAOC: Evaluating Agentic Ontology Curation of Drosophila Scientific Knowledge Bases
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2602.09163

- Nearest but Not Dearest: Shared Curator-Feedback Infrastructure for Content-Only Search and Recommendation
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30568

- KuaFu: Compressing Long User Behavior into Understanding at Billion Scale
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.31045

- Retail Product Search: A Practical Approach at Target
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.31498

- SignTrace: Describe a Sign, Find the Word
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30295

- High-probability guarantees for linear accessibility in feature superposition
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.09556

- QReason: Query-Focused Decoupled Chain-of-Thought for Efficient Passage Reranking
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30904

- Where Does Retrieval-Based Open-Ended Evaluation Fail? Automatic Taxonomy Induction from Long-Form Medical Answer Factuality Verification
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30467

- Bootstrapping Conversational Recommendation Agents At Spotify: Synthetic Data Generation and Self-Improvement Loops
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30297

- SPADE: Escaping the Popularity-Similarity Frontier to Measure Serendipitous Recommendations
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.31164

- RecToolBench: Benchmarking Recommendation-Specific Tool Orchestration under Fuzzy User Intent
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30717

- Component Benchmark: Hierarchical Model Profiling for Large-scale Recommendation Systems
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30656

- On Function-Correcting Codes in the Lee Metric
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2507.17654

- Enriching Sequential Recommendation with Graph Laplacian Positional Embeddings
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.31253

- Embedding Subspace Partitioning for Dynamic Multi-Objective Retrieval
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30601

- T-RoPE: Time-Aware Rotary Position Embedding for Sequential Recommendation
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30576

- Epstein Files Engine: Agentic Search for Investigative Journalism
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30611

- Recommendation World Models for Future-State Control
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30711

- REALMS: An AI-Assistant Conversational System for Real-Time Exact Audience Sizing over High-Dimensional Nested Profiles
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.30547

- CG-Probes: Recovering Guardrail Directions from Patient Query Embeddings
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2609.31062

- Bringing Agentic Search to Earth Observation Data Discovery
  Source: arxiv-cs-ir
  Topic: Information Retrieval (+AI, AI Research)
  URL: https://arxiv.org/abs/2607.02387

- Affective Flow Language Model for Emotional Support Conversation
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2602.08826

- Large Language Model Selection with Limited Annotations
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2605.24981

- Closing the Speech-Text Gap with Limited Audio for Effective Domain Adaptation in LLM-Based ASR
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2604.06487

- Intent2Tc: Automated Intent-to-Traffic Control Translation with Language Models
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31397

- AcuityBench: Evaluating Clinical Acuity Identification and Uncertainty Alignment
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2605.11398

- I-Parakeet: Integer-Only Conformer ASR on Mobile NPU
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30846

- Small Is Enough: Per-User Style Rewriting of AI-Edited Text via LoRA Adapters
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.29238

- A Survey on Fake Review Detection: From Pre-trained Language Models to Large Language Models
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30292

- Generating Legal Commentaries from Case Databases via Retrieval, Clustering, and Generation
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2605.24534

- From ASR to ASP: Evaluating Prompt Attack Vulnerabilities Against Open-Source LLMs
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2505.14368

- The Hard Part Comes After Search: Benchmarking Web Agents on Synthesizing, Organizing, and Displaying Knowledge
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30604

- UniPrefill: Universal Long-Context Prefill Acceleration via Block-wise Dynamic Sparsification
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2605.06221

- ToolSearcher: Optimizing Tool Selection at Scale via Reinforcement Learning
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30906

- Evidence-Grounded Auditing of Identification Assumptions in Climate-Policy Causal Evaluations
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30867

- LAVOIR: Teaching a Single-Pass Decision Encoder When and What to Ask with Amortized Value of Information
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30706

- Rufus-Air: An Open LLM Post-Training Recipe
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.29421

- User Model Extraction via Belief Self-Distillation
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31603

- Don't CLAP: Are Music-Text Models Bag-of-Words?
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30540

- The Right Information Extraction Pipeline Depends on the Document: Accuracy-Energy Trade-offs for Small, Local Models
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31341

- Compact Documentation for Coding Agents: A Benchmark, an Optimizer, and Why It Does Not Transfer
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31587

- Apollo Restore: A Foundation LLM for Historical Greek Optimized for Fill-in-the-Middle Restoration of Ancient Greek Texts
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.22455

- PALM: Point-in-Time Adaptation for Financial Language Models
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30316

- TRACE: Temporal Audit and Condition-aware Evaluation of Streaming Video Understanding
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30670

- G$^2$PTQ: Improving LLM Post-Training Quantization with Generalized Gradient Compensation
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31009

- Evaluating Cultural Awareness of LLMs for Haitian Creole
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31506

- Highlight-Then-Summarize: Learning to Compress Evidence for Long-Context Understanding
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31382

- Not All Memories Are Equal: Hierarchical Collaborative Memory for Validity-Aware Retrieval in LLM Agents
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30289

- COT-TTS: Audio Context-Aware Text-to-Speech with Chain-of-Thought Reasoning
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.22697

- Words Speak Louder Than Order: A Behavioral Evaluation of Gemma 4
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30716

- Attention-Discounted Adaptive Sampler for Masked Diffusion Language Models
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.10829

- Layer-wise Target Propagation: Efficient Component Attribution through Target Centric Propagation
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.19742

- Inference-Time Target Speaker Unlearning in LLM-Based Automatic Speech Recognition
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30439

- Does Uniform Discrete Diffusion Need Time?
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30977

- Achieving Tokenizer Flexibility in Language Models through Heuristic Adaptation and Supertoken Learning
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2505.09738

- Cross-Backend QIEO: Universal Runtime Portability across OpenMP5, CUDA, HIP, and Multi-Language Interfaces
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30914

- Towards Automated Lexicography: Generating and Evaluating Definitions for Learner's Dictionaries
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2601.01842

- CARGO: Context-Aware Retrieval-Gated Evaluation of Agentic AI in Production
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30471

- Manifold Projection and Iterative Autoencoder Refinement for Masked Language Modeling
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30288

- SlideLab: Audience-Centered Scientific Slide Generation and Evaluation
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30294

- Thinking Less to Simulate Better: Intuitive Prompting Improves LLM Agents Simulating Individual Social Media Reactions, Including Unfamiliar Content
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30563

- Prompt Injection Detection for Email Agents Through Attack Chain Modeling
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30657

- Modeling Student Sensemaking with LLMs and Knowledge-Graph-Guided Inference
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31046

- Why Alzheimer's Speech Screening Fails to Generalize: Bridging the Deployment Gap via Cross-Corpus Evidence Anchoring
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31293

- PUBG Ally: A Conversational Embodied Agent as an AI Teammate
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.29837

- State of Thought Enables Endogenous Reasoning
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.16055

- Depth-adaptive Inference of Looped Language Models via Continuous Depth Batching
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.09444

- Stale-Document Poisoning: When Outdated Retrieval Overrides Correct Model Answers
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31342

- Symbiotic Architecture for Post-Hoc Audio Extension of Frozen Language Models
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30784

- Inquesto Score: A reliability Protocol For Voice Agents
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30514

- Rethinking Human-Aligned Evaluation: An Analysis of Semantic Metrics Beyond WER
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.21663

- The Communication Map of a Transformer
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.22007

- NaijaNLP: A Survey of Nigerian Low-Resource Languages
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2502.19784

- PriceBench: A Diagnostic Benchmark for Price, Quality, and Brand Preferences in LLM Booking Agents
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31468

- SkillFlow: Scalable and Efficient Agent Skill Retrieval System
  Source: arxiv-cs-cl
  Topic: AI Research (+AI, Information Retrieval)
  URL: https://arxiv.org/abs/2504.06188

- AcoustiClaim: A Numeric Claim Benchmark with Instrument Ground Truth
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30483

- GT-HarmBench: Benchmarking AI Safety Risks Through the Lens of Game Theory
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2602.12316

- Statistical Foundations for a Google Play User-Review Sentiment Index: Signal Fusion, Shrinkage, Distributional Validation, and Dynamic Smoothing
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31513

- Understanding the Role of Prompt Template in Knowledge Distillation for Safety Alignment
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30802

- All In Good Time: Causality-Aware Framework for LLM-Based Simultaneous Speech-to-Speech Translation
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30416

- Asymmetric Classifier-Free Guidance for Target-Speaker ASR
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30476

- Beyond Mean Attention: Diversity-Aware, Layer-Wise Scoring for KV Cache Eviction
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30738

- Learning Natural Conversational Behavior in Tandem Speech-to-Speech Models with Randomized Guidance
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30773

- Feeding BabyLMs Macaroni: Code-Switching Curricula Cause Cross-Lingual Convergence
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30535

- Coupled Usage-Sense Processes: Temporal and Attributable Lexical Semantic Change
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30974

- INFUSER: Influence-Guided Self-Evolution Improves Reasoning
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.09052

- Muslim: A Deployed Arabic Voice AI Platform for Grounded Islamic Knowledge
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31511

- THA: Weighted Finite-State Text Normalization and Inverse Text Normalization for Khmer
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30984

- Likelihood Ranking doesn't Scale Like Prompting in LLMs
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.29390

- VLAA-GUI: Knowing When to Stop, Recover, and Search, A Modular Framework for GUI Automation
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2604.21375

- Stepwise Intrinsic Rewards for Reasoning in Large Language Models
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2602.01034

- MedHal: a Synthetic Dataset for Medical Hallucination Detection
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2504.08596

- When Is a Multi-Agent Code Judge Actually Grounded? Two Label-Free Measurements, and a Judge That Declines to Guess
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30328

- Calibrated Enough to Know, Not Calibrated to Act: Fabricated Evidence Makes LLM Agents Commit to the Unknowable
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.27167

- Persistent Negatives for Adversarial Black-Box On-Policy Distillation
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30864

- Recursive Self-Improvement via On-Policy Distillation for Reasoning
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30652

- Asking For An Old Friend: Diagnosing and Mitigating Temporal Failure Modes in LLM-based Statutory Question Answering
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2605.23497

- A Mechanistic Study of AI-Text Detection Neurons in Frozen BERT: Sparse Probing and Activation Patching on RAID
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30287

- Beyond Atomic Tokens: Factorizing Syllables for Language Model Pretraining
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.21362

- Identifying Scientists on X
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31264

- JevAdvBench: A Benchmark and Black-Box Attacks for Reinforcement Learning for Calibrated Decisions Models
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31142

- Sorry Robot, Happy Human: Vision-Language Models Read Only One of Two Legible Typographic Layers
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31403

- Same Text, Different Numbers: The Divergence of LLM-Based Measures
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31013

- MexHat: A Dataset for Hate Speech Detection in Mexican Spanish Videos
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31553

- Enhancing Assessment of Self-Consistency in LLM Explanations using Perturbation Strength
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30849

- Estimating and Orthogonalizing Unknown Pre-training Gradients for Continual Fine-tuning of Large Language Models
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30935

- Evaluation is All You Need: Strategic Overclaiming of LLM Reasoning Capabilities Through Evaluation Design
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2506.04734

- Auditing and Repairing LLM-as-Judge Failures in a Production Text-to-SQL Pipeline
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30290

- LocUS: Head Selection and Subspace Projection for Targeted Activation Steering
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31122

- A Benchmark Framework for Screening Automation in Systematic Reviews
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30298

- Learning to Stop without Learning to Stop: Self-Supervised Confidence Training Improves Reasoning Efficiency
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31619

- Evaluating Sycophancy in Chinese Large Language Models on Factual Questions Derived from Online Search Queries
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30986

- Strategically Diverse Sampling for Self-Training
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31571

- Does Understanding Inform Generation in Unified Multimodal Models? From Analysis to Path Forward
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2511.20561

- SEA-CLIP-Tiny: Efficient Multilingual Text-Vision Embedding for Southeast Asian Languages
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30739

- Two Conformal Constructions for Adaptive Within-Document AI-Text Screening
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31547

- Towards Mitigating Fabricated Consensus: The Active Provenance Gate for Multi-Agent Debate Synthesis
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31422

- Where a Model Sends Its Own Repeated Token
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31181

- What Improves Multimodal Misinformation Detection? Answers from a Large-Scale Empirical Study
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30402

- Human-1 by Josh Talks: A Full-Duplex Conversational Modeling Framework in Hindi using Real-World Conversations
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2604.23295

- ZooWork-ShopRanker: An Open, Preference-Aligned E-Commerce Reranker
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31002

- Training-Free Pronunciation Transcription via Text-Constrained Acoustic Rescoring
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30924

- ViSTA: A Simple Bridge Extends Visual Alignment to Clinical Time-Series Understanding in Multimodal LLMs
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31448

- ArGuard Shared Task: Harmful Content Detection in Arabic Memes and LLM Prompts
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.29349

- Combee: Scaling Prompt Learning for Self-Improving Language Model Agents
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2604.04247

- PrivDrift: Auditing User-Secret Leakage Under Topic Drift in Active LLM Conversations
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30094

- Improving Visual Sensitivity of LLMs on Multimodal Machine Translation with Metric-based Loss Weighting
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31169

- Probing Stability-Plasticity Tradeoffs in Agent Memory through Cognitive Experimental Paradigms
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30558

- RupeeBias: Auditing Demographic Bias in Indian Economic Guidance from Large Language Models
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31245

- LeakScale: Estimating the Causal Effect of Benchmark Exposure
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.27176

- RAZOR: Pruning Replaceable Experts in LLMs
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30465

- Direct Preference Optimization for English-Mandarin Code-Switching Speech Recognition in Audio LLMs
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2605.23975

- StraTA: Incentivizing Agentic Reinforcement Learning with Strategic Trajectory Abstraction
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2605.06642

- FAVoR: Measuring and Mitigating Author-Style Homogenization in Federated Personalized Generation
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30968

- Quantizing Looped Transformers: Feedback Exposure and Calibration Blindness
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30820

- Do we need to answer that question? Salience and Answerability of Potential Questions in Naturalistic Dialogue
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31130

- Why Better Cross-Lingual Alignment Fails for Better Cross-Lingual Transfer: Case of Encoders
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.18863

- PIA: A Personal Intelligence Agent Turning Health Conversations into Records and Records into Understanding
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31255

- Breaking Homogeneity: Diversifying Persona Sets for Creative LLM Outputs
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30492

- Cartograph: Federated Tool Discovery with Operator-Attested Retrieval for AI Agents
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30293

- A Unified Account of Concepts and Chunks
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30414

- Statistical Priors for Implicit Preferences: Decoupling Skill Selection as a Local Harness in Personal Agents
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.05828

- From annotation to reasoning: Culture in language models
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30897

- MoSAR: Mixture of Semantic Attention Regimes for Learning Adaptive and Approximable Attention Geometries
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31261

- Effects of Transcript Compression on LLM-based Medical Misinformation Detection in Japanese YouTube Videos
  Source: arxiv-cs-cl
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30882

- UQ-LOB: Uncertainty-Aware Limit Order Book Mid-Price Forecasting
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31491

- MOPD-Router: Rethinking Teacher Routing in Multi-Teacher On-Policy Distillation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30837

- Analyzing and Mitigating Cost-Inefficient Behaviors in Coding Agents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30725

- Coding Agents Aren't Enough! Evaluating an Enterprise Security Brain for Agentic Cloud Investigations
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30345

- Softmax Reparameterization for Output-Head Quantization
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31291

- Cognitive Skills in the Age of AI: Computing Students and Experts Perceptions
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31272

- Robust to Which Model Change? A Unified Evaluation of Robust Counterfactual Explanations
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.30918

- UniAR: A Unified Framework for Autism Recognition Enhanced by Multi-View Prompt Learning
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.31298

- Not Every Divergence Should Be Suppressed: Counterfactual Recoverability in On-Policy Distillation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.04408