Copertina

When Benchmarks Leak: Inference-Time Decontamination for LLMs

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Chai, Jianzhe, Zhe, Yu, Sakuma, Jun
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Computation and Language Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

One Leak Away: How Pretrained Model Exposure Amplifies Jailbreak Risks in Finetuned LLMs
di: Tan, Yixin, et al.
Pubblicazione: (2025)

Inference-Time Decontamination: Reusing Leaked Benchmarks for Large Language Model Evaluation
di: Zhu, Qin, et al.
Pubblicazione: (2024)

Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMs
di: Balloccu, Simone, et al.
Pubblicazione: (2024)

When LLMs Benchmark Themselves: Deconstructing Self-Bias in Automated Evaluation
di: Xu, Wenda, et al.
Pubblicazione: (2025)

LessLeak-Bench: A First Investigation of Data Leakage in LLMs Across 83 Software Engineering Benchmarks
di: Zhou, Xin, et al.
Pubblicazione: (2025)

When Personalization Misleads: Understanding and Mitigating Hallucinations in Personalized LLMs
di: Sun, Zhongxiang, et al.
Pubblicazione: (2026)

PAD: Personalized Alignment of LLMs at Decoding-Time
di: Chen, Ruizhe, et al.
Pubblicazione: (2024)

Agent Tools Orchestration Leaks More: Dataset, Benchmark, and Mitigation
di: Qiao, Yuxuan, et al.
Pubblicazione: (2025)

Bounded Rationality for LLMs: Satisficing Alignment at Inference-Time
di: Chehade, Mohamad, et al.
Pubblicazione: (2025)

When Life Gives You Samples: The Benefits of Scaling up Inference Compute for Multilingual LLMs
di: Khairi, Ammar, et al.
Pubblicazione: (2025)

Regression-aware Inference with LLMs
di: Lukasik, Michal, et al.
Pubblicazione: (2024)

DeepMath-103K: A Large-Scale, Challenging, Decontaminated, and Verifiable Mathematical Dataset for Advancing Reasoning
di: He, Zhiwei, et al.
Pubblicazione: (2025)

Improving Factuality in LLMs via Inference-Time Knowledge Graph Construction
di: Wu, Shanglin, et al.
Pubblicazione: (2025)

Summarize Before You Speak with ARACH: A Training-Free Inference-Time Plug-In for Enhancing LLMs via Global Attention Reallocation
di: Wang, Jingtao, et al.
Pubblicazione: (2026)

Knowing When to Abstain: Medical LLMs Under Clinical Uncertainty
di: Machcha, Sravanthi, et al.
Pubblicazione: (2026)

Benchmarking and Adapting On-Device LLMs for Clinical Decision Support
di: Munim, Alif, et al.
Pubblicazione: (2025)

Leaking LoRa: An Evaluation of Password Leaks and Knowledge Storage in Large Language Models
di: Marinelli, Ryan, et al.
Pubblicazione: (2025)

LoRA-Leak: Membership Inference Attacks Against LoRA Fine-tuned Language Models
di: Ran, Delong, et al.
Pubblicazione: (2025)

DHP Benchmark: Are LLMs Good NLG Evaluators?
di: Wang, Yicheng, et al.
Pubblicazione: (2024)

Humans and LLMs Diverge on Probabilistic Inferences
di: Kamath, Gaurav, et al.
Pubblicazione: (2026)

Tandem Transformers for Inference Efficient LLMs
di: S, Aishwarya P, et al.
Pubblicazione: (2024)

Draft-based Approximate Inference for LLMs
di: Galim, Kevin, et al.
Pubblicazione: (2025)

LLMs Are Prone to Fallacies in Causal Inference
di: Joshi, Nitish, et al.
Pubblicazione: (2024)

Say Anything but This: When Tokenizer Betrays Reasoning in LLMs
di: Ayoobi, Navid, et al.
Pubblicazione: (2026)

Generative Adversarial Reviews: When LLMs Become the Critic
di: Bougie, Nicolas, et al.
Pubblicazione: (2024)

Keep Guessing? When Considering Inference Scaling, Mind the Baselines
di: Yona, Gal, et al.
Pubblicazione: (2024)

LLMEval-Med: A Real-world Clinical Benchmark for Medical LLMs with Physician Validation
di: Zhang, Ming, et al.
Pubblicazione: (2025)

Beyond Benchmark: LLMs Evaluation with an Anthropomorphic and Value-oriented Roadmap
di: Wang, Jun, et al.
Pubblicazione: (2025)

Fairness Evaluation and Inference Level Mitigation in LLMs
di: Nadeem, Afrozah, et al.
Pubblicazione: (2025)

POLYCHARTQA: Benchmarking Large Vision-Language Models with Multilingual Chart Question Answering
di: Xu, Yichen, et al.
Pubblicazione: (2025)

ConsistencyAI: A Benchmark to Assess LLMs' Factual Consistency When Responding to Different Demographic Groups
di: Banyas, Peter, et al.
Pubblicazione: (2025)

When LLMs Team Up: The Emergence of Collaborative Affective Computing
di: Lai, Wenna, et al.
Pubblicazione: (2025)

Surgical Feature-Space Decomposition of LLMs: Why, When and How?
di: Chavan, Arnav, et al.
Pubblicazione: (2024)

DICE: Structured Reasoning in LLMs through SLM-Guided Chain-of-Thought Correction
di: Li, Yiqi, et al.
Pubblicazione: (2025)

POPI: Personalizing LLMs via Optimized Natural Language Preference Inference
di: Chen, Yizhuo, et al.
Pubblicazione: (2025)

Can LLMs Correct Themselves? A Benchmark of Self-Correction in LLMs
di: Tie, Guiyao, et al.
Pubblicazione: (2025)

LLMs Should Incorporate Explicit Mechanisms for Human Empathy
di: You, Xiaoxing, et al.
Pubblicazione: (2026)

Pattern Enhanced Multi-Turn Jailbreaking: Exploiting Structural Vulnerabilities in Large Language Models
di: Nihal, Ragib Amin, et al.
Pubblicazione: (2025)

Detecting Hallucinations in SpeechLLMs at Inference Time Using Attention Maps
di: Waldendorf, Jonas, et al.
Pubblicazione: (2026)

Activation-aware Probe-Query: Effective Key-Value Retrieval for Long-Context LLMs Inference
di: Xiao, Qingfa, et al.
Pubblicazione: (2025)