Language Model Auditing

Follow this topic Ask about this topic

Momentum

22 papers in the last four weeks, up 100% on the four weeks before. 0.2% of all new papers.

Jul 13Week of Sep 28

Latest papers 110

All topics

Similarity Gates Approve Reversals: A Validity Audit of Embedding-Cosine Thresholds in Agent Systems
Aug 10, 2026Scott E. FriasSemantic Textual Similarity Text Embedding Evaluation
Eigenforma · Freemind Labs
Measuring the Tokenization Premium: A Cost Audit for Underserved Language Communities
Aug 10, 2026Avijit Roy, Proma Roy, Hrishitva PatelAlgorithmic Fairness Token Efficiency
John Jay College of Criminal Justice, City University of New York, USA · The City College of New York, City University of New York, USA · University of Texas at San Antonio, USA
Mind the Hook: Source-Level Auditing of Privacy Defenses in Retrieval-Augmented Generation
Aug 10, 2026Yanhang Li, Zhichao Fan, Zexin ZhuangRetrieval-Augmented Generation RAG Security
Northeastern University · University of Illinois Urbana-Champaign · Southern Methodist University
Do Audio Language Models Use Paralinguistic Evidence? Counterfactual Audits for Response Evaluation
Aug 7, 2026Kevin Miller, Arjun Chandra, Venkatesh SaligramaAudio-Language Model Evaluation Language Model Auditing
Boston University
Why AI Detection Fails for Academic Integrity
Aug 6, 2026Jonathan A. Karr, Grigorii Khvatskii, Ting Hua +1AI-Generated Text Detection Language Model Auditing
University of Notre Dame Notre Dame, Indiana, USA
Signal or Spurious Cue? A Randomized Audit of Survey-Country Metadata in LLM Social Inference
Aug 6, 2026Yifan Lyu, Xinran Li, Jiaqi Qiao +1Probabilistic Forecasting Causal Interventions in Language Models
Dalian University of Technology, China
Auditing Data Provenance in LLM Fine-tuning via Intrinsic Distributional Fingerprints
Aug 3, 2026Zirui Huang, Yunlong Mao, Wei Tong +3Language Model Fingerprinting Training Data Attribution
State Key Laboratory of Novel Software Technology, Nanjing University Nanjing, Jiangsu, China
Auditing Medical Vision-Language Models on Chest Radiographs: Estimating Reference Agreement Across Institutions
Aug 1, 2026Pengyang Yu, Yiou Wang, Zhongping Dong +3Chest X-Ray Classification Medical VLMs
School of Computer Science, University College Dublin, Dublin, Ireland · Department of Medical Imaging, The Third Affiliated Hospital of Southern Medical University, Guangzhou, China · Dublin City University, Dublin, Ireland
AIriskEval-edu Demo: Auditing of Pedagogical Risks in Educational Explanations
Jul 28, 2026Javier Irigoyen, Roberto Daza, Francisco Jurado +5Generative AI in Education Explainability Evaluation
BiometricsAI, Universidad Autónoma de Madrid (UAM), Spain · GHIA, Universidad Autónoma de Madrid (UAM), Spain · Universidad de Las Palmas de Gran Canaria (ULPGC), Spain
Who Gets Named: Citation Type Predicts Individual Naming by Grounded Language Models, and a Roster Instrument Captures 0.5% of It
Jul 26, 2026Dmitrij ŻatuchinGenerative AI Evaluation Language Model Auditing
Rankfor.AI OÜ, Tallinn · Rankfor.AI OÜ, Tallinn, Estonia
When Activation Oracles Learn Not to Read: Concept-Specific Blind Spots in Fine-Tuned Oracles
Jul 25, 2026Tobias Bersia, Tatiana GaintsevaLLM Interpretability Language Model Auditing
BAISH · Queen Mary University of London
Learning to Detect UI Principle Violations via Reinforcement Learning
Jul 22, 2026Nishi Mehta, Swathi Alse, Himani Kumawat +2Automated Evaluation Language Model Auditing
University of California, Santa Cruz · Carnegie Mellon University
Train the Model, Not the Reader: Decodability Supervision for Verifiable Activation Explanations
Jul 22, 2026Hiskias DingetoFaithfulness of Language Model Explanations LLM Interpretability
StackOne Technologies
Measuring Reward-Seeking via Contrastive Belief Updates
Jul 21, 2026Axel Højmark, Jérémy Scheurer, Evgenia Nitishinskaya +5RL for Language Models LLM Alignment
Apollo Research · OpenAI
Traccia: An OpenTelemetry-Based Governance Platform for AI Systems
Jul 15, 2026Nutan Kumar Naik, Aditya Kumar Saroj, Vijay Prasad Poudel +2AI Accountability AI Agent Monitoring
Department of Earth and Atmospheric Sciences National Institute of Technology Rourkela Odisha, India · Algen.AI2026 Bengaluru, Karnataka, India
What Models Express, Suppress, and Resist: Auditing Open-Weight LLMs with Persona Vectors
Jul 14, 2026Winston Zeng, Ali Emami, Jinho ChoiLanguage Model Steering LLM Interpretability
Emory University
TRACE: An Operational Reasoning Schema for Auditable Agentic Commitments
Jul 14, 2026Edward Y. Chang, Emily J. ChangData Provenance Reasoning Evaluation
Stanford University · Quadrivium AI
The Paternalistic Filter: Epistemic Injustice and Differential Refusal in LLM-Mediated History Education for Marginalized Romanian Students
Jul 13, 2026Alexis Popovici, Andrei Ionascu, Adrian-Marius DumitranLanguage Model Safety Evaluation AI in Education
Universitatea din Bucuresti, Bucharest, Romania
The Entanglement Wall: Activation-Space Probes as Risk Detectors, Not Context Adjudicators
Jul 12, 2026Dominik SchwarzLanguage Model Safety Evaluation Hidden-State Probing
Independent Researcher
Creativity, honesty and designed forgetting emerge in small hyperbolic language models
Jul 10, 2026Kwan Soo Shin, In Seok Kang, Yunkyung MinLLM Sycophancy Memory-Augmented Language Models
PolymathMinds Lab, Seoul, Republic of Korea. · Department of Chemical Engineering, POSTECH, Pohang, and aSSIST University, Seoul, Republic of Korea. · Korean Educational Development Institute, Jincheon, Republic of Korea.
NL-PAC: Specification Ambiguity and Certified Minimax Risk Floors in LLM-Mediated Supervision
Jul 9, 2026Berkay AnahtarciLLM Evaluation Partial Identification
Department of Mathematical Engineering, Özyeğin University, Istanbul, Türkiye
Token-Flow Firewall: Semantic Runtime Auditing for Persistent AI Agents
Jul 9, 2026Puji Wang, Yingchen Zhang, Ruqing Zhang +2AI Agent Security AI Agent Monitoring
State Key Laboratory of AI Safety · Institute of Computing Technology, Chinese Academy of Sciences · University of Chinese Academy of Sciences, Beijing, China
Beyond Static Rules: Automated Discovery of Latent Vulnerabilities in Text-to-SQL
Jul 4, 2026Hanqing Wang, Yongdong Chi, Jian Yang +4LLM Security Text-to-SQL
Shanghai University of Finance and Economics · Beihang University · Deepexi Technology Co. Ltd. +1
When Aggregate Alignment Misleads: Auditing Policy Repair Without Per-State Expert Actions
Jul 3, 2026Peiying Zhu, Sidi ChangLLM Agent Evaluation Policy Evaluation
Blossom AI San Francisco, USA · Blossom AI Labs Tokyo, Japan
Retrieval Observability Bounds on Provenance Detection for Agent Memory Poisoning: Measured Coverage and a Falsified Standalone Detector
Jun 29, 2026Jun Wen LeongRAG Poisoning Attacks Agent Memory Poisoning
EMPATH: A Multilingual Auditor-Judge Benchmark for Safety Evaluation of Emotional-Support Chatbots
Jun 29, 2026Camilo Chacón SartoriLLM Safety Benchmarks Multilingual Language Model Evaluation
MindSurf
Defeat Devices in AI Systems
Jun 27, 2026Emilio FerraraDeception in Language Models AI Safety Evaluation
Thomas Lord Department of Computer Science University of Southern California
Black-Box Forensics for Conversational LLM Agents
Jun 21, 2026Isadora White, Yasaman Jafari, Taylor Berg-KirkpatrickLanguage Model Fingerprinting LLM Agent Security
University of California, San Diego
Channel Location Constrains the Auditability of Subliminal Learning
Jun 20, 2026Tamas MadlSubliminal Learning Language Model Distillation
Austrian Research Institute for Artificial Intelligence · University of Vienna
Towards Understanding and Measuring COGNITIVE ATROPHY in LLM Behaviour
Jun 16, 2026Abeer Badawi, Moyosoreoluwa Olatosi, Negin Baghbanzadeh +5Emotional Support Conversation Mental Health
York University, Canada · Vector Institute, Canada · Dalhousie University, Canada +3

Aug 10, 2026, cs.CLJ/K move · Enter open · S save

Similarity Gates Approve Reversals: A Validity Audit of Embedding-Cosine Thresholds in Agent Systems

Scott E. Frias

Eigenforma · Freemind Labs

Semantic Textual Similarity Text Embedding Evaluation Agent Evaluation Semantic Change Detection Language Model Auditing

Open paper AskPDF arXiv

Agent frameworks ship quality gates that compare text blocks by embedding-cosine similarity and decide at a fixed cutoff. Deduplication filters, semantic caches, drift guards, and answer grader gates deploy to answer the question: "Does this text still mean the same thing?" But the score answers a different question: "How much did the wording change?" We audit this gate class as a measurement instrument. In the cases these gates exist to catch, the two can run in opposite ways. Many times, reversing an instruction is a single word edit, while agreement often rephrases a sentence. The consequence is a safety check that fires backwards. The production drift guard we audited caught 0 of 56 meaning-breaking mutations, and one approved item, "withhold the study drug" -> "administer the study drug", came in at cosine 0.9608. We observed five shipped operating points, and balanced accuracy across 90 configuration-threshold-task cells never exceeded 0.700 (median 0.525). The same confounder also corrupted evaluations. A naively built corpus inherits this confounder and can return an inverted verdict, with a decision AUROC exactly 0.000 in 13 of 18 configuration-task cells (at most 0.040 in all 18) against 0.440-0.815 for the same nine configurations under a balanced 2x2 design. Twice in the effort it captured our own headline claims. Obvious repairs fail: an encoder swap and an overlap-conditioned gate (0.750 in-sample, 0.533 held-out) land at chance on separately authored held-out data, and an NLI drop-in did no better. Embeddings do still bear hope here, as the strongest two of nine configurations separated reversal from paraphrase at matched overlap (AUROC 0.79-0.90), but only a matched-pair audit reveals the deployment regime. We release the corpus method, harness, and frozen results, and contend that scores gated this way measure the wrong thing. We believe a valid instrument is buildable.

Momentum

Related topics

Latest papers 110