Which Model Is Actually Serving You? IRIS: Budgeted Black-Box Auditing of Model Substitution and Routing Dilution in LLM Gateways
Yuewei Zhang, Zhi-Hai Zhang, Hanzhang Qin
cs.CR
Submitted: 2026-07-23
Code: https://github.com/Photen/IRIS-audit
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- LLMs and the Madness of Crowds
- Are You Getting What You Pay For? Auditing Model Substitution in LLM APIs
- KBF: Knowledge Boundary as Fingerprint for Language Model and Black-Box API Auditing
- Incompressible Knowledge Probes: Estimating Black-Box LLM Parameter Counts via Factual Capacity
- Your Agent Is Mine: Measuring Malicious Intermediary Attacks on the LLM Supply Chain
- LLMLagBench: Identifying Temporal Training Boundaries in Large Language Models
- Your "Pro" LLM Subscription May Actually Be "Free": Exposing Fingerprint Spoofing Risks in LLM Inference Services
- Real Money, Fake Models: Deceptive Model Claims in Shadow APIs
- One Token Is Enough: Fingerprinting and Verifying Large Language Models from Single-Token Output Distributions
- You've Changed: Detecting Modification of Black-Box Large Language Models
- TOPLOC: A Locality Sensitive Hashing Scheme for Trustless Verifiable Inference
Related papers
- SoK: AI-Augmented Binary Reversing
- Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
- Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
- Efficient Fuzzy PSI under One-Sided Assumptions
- Sealing the Audit-Runtime Gap for LLM Skills
- Token Composition: A Graph Based on EVM Logs