Cover Image

DyePack: Provably Flagging Test Set Contamination in LLMs Using Backdoors

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Cheng, Yize, Wang, Wenxiao, Moayeri, Mazda, Feizi, Soheil
Format:	Preprint
Published:	2025
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

SpurLens: Automatic Detection of Spurious Cues in Multimodal LLMs
by: Hosseini, Parsa, et al.
Published: (2025)

PRIME: Prioritizing Interpretability in Failure Mode Extraction
by: Rezaei, Keivan, et al.
Published: (2023)

Tool Preferences in Agentic LLMs are Unreliable
by: Faghih, Kazem, et al.
Published: (2025)

Chain-of-Defensive-Thought: Structured Reasoning Elicits Robustness in Large Language Models against Reference Corruption
by: Wang, Wenxiao, et al.
Published: (2025)

Maestro: Joint Graph & Config Optimization for Reliable AI Agents
by: Wang, Wenxiao, et al.
Published: (2025)

Your LLM Agents are Temporally Blind: The Misalignment Between Tool Use Decisions and Human Time Perception
by: Cheng, Yize, et al.
Published: (2025)

Adversarial Paraphrasing: A Universal Attack for Humanizing AI-Generated Text
by: Cheng, Yize, et al.
Published: (2025)

Schoenfeld's Anatomy of Mathematical Reasoning by Language Models
by: Li, Ming, et al.
Published: (2025)

Off-the-Shelf LLMs as Process Scorers: Training-Free Alternative to PRMs for Mathematical Reasoning
by: Chegini, Atoosa, et al.
Published: (2026)

Improving Compositional Attribute Binding in Text-to-Image Generative Models via Enhanced Text Embeddings
by: Zarei, Arman, et al.
Published: (2024)

Rethinking Artistic Copyright Infringements in the Era of Text-to-Image Generative Models
by: Moayeri, Mazda, et al.
Published: (2024)

Can AI-Generated Text be Reliably Detected?
by: Sadasivan, Vinu Sankar, et al.
Published: (2023)

Almost AI, Almost Human: The Challenge of Detecting AI-Polished Writing
by: Saha, Shoumik, et al.
Published: (2025)

SpecHop: Continuous Speculation for Accelerating Multi-Hop Retrieval Agents
by: Saberi, Mehrdad, et al.
Published: (2026)

Endor: Hardware-Friendly Sparse Format for Offloaded LLM Inference
by: Joo, Donghyeon, et al.
Published: (2024)

Unearthing Skill-Level Insights for Understanding Trade-Offs of Foundation Models
by: Moayeri, Mazda, et al.
Published: (2024)

Reasoning Under Uncertainty: Exploring Probabilistic Reasoning Capabilities of LLMs
by: Pournemat, Mobina, et al.
Published: (2025)

Embracing Diversity: Interpretable Zero-shot classification beyond one vector per class
by: Moayeri, Mazda, et al.
Published: (2024)

Quantifying the Effect of Test Set Contamination on Generative Evaluations
by: Schaeffer, Rylan, et al.
Published: (2026)

A Closer Look at Bias and Chain-of-Thought Faithfulness of Large (Vision) Language Models
by: Balasubramanian, Sriram, et al.
Published: (2025)

Evaluation of LLMs in Speech is Often Flawed: Test Set Contamination in Large Language Models for Speech Recognition
by: Tseng, Yuan, et al.
Published: (2025)

Pack of LLMs: Model Fusion at Test-Time via Perplexity Optimization
by: Mavromatis, Costas, et al.
Published: (2024)

DCR: Quantifying Data Contamination in LLMs Evaluation
by: Xu, Cheng, et al.
Published: (2025)

RESTOR: Knowledge Recovery in Machine Unlearning
by: Rezaei, Keivan, et al.
Published: (2024)

Early Stopping for Large Reasoning Models via Confidence Dynamics
by: Hosseini, Parsa, et al.
Published: (2026)

On Mechanistic Circuits for Extractive Question-Answering
by: Basu, Samyadeep, et al.
Published: (2025)

Benchmarking ChatGPT and DeepSeek in April 2025: A Novel Dual Perspective Sentiment Analysis Using Lexicon-Based and Deep Learning Approaches
by: Alhusseini, Maryam Mahdi, et al.
Published: (2025)

Pruning Strategies for Backdoor Defense in LLMs
by: Chapagain, Santosh, et al.
Published: (2025)

MetaBackdoor: Exploiting Positional Encoding as a Backdoor Attack Surface in LLMs
by: Wen, Rui, et al.
Published: (2026)

Let the Bees Find the Weak Spots: A Path Planning Perspective on Multi-Turn Jailbreak Attacks against LLMs
by: Liu, Yize, et al.
Published: (2025)

Quantifying Data Contamination in Psychometric Evaluations of LLMs
by: Han, Jongwook, et al.
Published: (2025)

Fast Adversarial Attacks on Language Models In One GPU Minute
by: Sadasivan, Vinu Sankar, et al.
Published: (2024)

MedRedFlag: Investigating how LLMs Redirect Misconceptions in Real-World Health Communication
by: Sambara, Sraavya, et al.
Published: (2026)

Decomposition-Enhanced Training for Post-Hoc Attributions In Language Models
by: Balasubramanian, Sriram, et al.
Published: (2025)

Task-Aware Calibration: Provably Optimal Decoding in LLMs
by: Tomov, Tim, et al.
Published: (2026)

Evaluate Bias without Manual Test Sets: A Concept Representation Perspective for LLMs
by: Gao, Lang, et al.
Published: (2025)

Geometry of Semantics in Next-Token Prediction: How Optimization Implicitly Organizes Linguistic Representations
by: Zhao, Yize, et al.
Published: (2025)

Explain the Flag: Contextualizing Hate Speech Beyond Censorship
by: Liartis, Jason, et al.
Published: (2026)

Explore the Reasoning Capability of LLMs in the Chess Testbed
by: Wang, Shu, et al.
Published: (2024)

Threshold Filtering Packing for Supervised Fine-Tuning: Training Related Samples within Packs
by: Dong, Jiancheng, et al.
Published: (2024)