Copertina

LaRA: Layer-wise Representation Analysis for Detecting Data Contamination in RL Post-Training

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Gwak, Minju, Kwak, Minseo, Lee, Dongseok, Son, Guijin, Ritter, Alan, Kim, Jaehyung
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Machine Learning Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Revisiting the Uniform Information Density Hypothesis in LLM Reasoning
di: Gwak, Minju, et al.
Pubblicazione: (2025)

Revisiting the UID Hypothesis in LLM Reasoning Traces
di: Gwak, Minju, et al.
Pubblicazione: (2025)

Gap-K%: Measuring Top-1 Prediction Gap for Detecting Pretraining Data
di: Kwak, Minseo, et al.
Pubblicazione: (2026)

Training-free LLM Verification via Recycling Few-shot Examples
di: Lee, Dongseok, et al.
Pubblicazione: (2025)

LaRA: Benchmarking Retrieval-Augmented Generation and Long-Context LLMs -- No Silver Bullet for LC or RAG Routing
di: Li, Kuan, et al.
Pubblicazione: (2025)

The Amazing Agent Race: Strong Tool Users, Weak Navigators
di: Kim, Zae Myung, et al.
Pubblicazione: (2026)

ResearchMath-14K: Scaling Research-Level Mathematics via Agents
di: Son, Guijin, et al.
Pubblicazione: (2026)

TiTok: Transfer Token-level Knowledge via Contrastive Excess to Transplant LoRA
di: Jung, Chanjoo, et al.
Pubblicazione: (2025)

OmniDrop: Layer-wise Token Pruning for Omni-modal LLMs via Query-Guidance
di: Park, Yeo Jeong, et al.
Pubblicazione: (2026)

Anomaly Detection with Adaptive and Aggressive Rejection for Contaminated Training Data
di: Lee, Jungi, et al.
Pubblicazione: (2025)

AlphaFree: Recommendation Free from Users, IDs, and GNNs
di: Jeon, Minseo, et al.
Pubblicazione: (2026)

An intuitive multi-frequency feature representation for SO(3)-equivariant networks
di: Son, Dongwon, et al.
Pubblicazione: (2024)

DEF-oriCORN: efficient 3D scene understanding for robust language-directed manipulation without demonstrations
di: Son, Dongwon, et al.
Pubblicazione: (2024)

FlexRound: Learnable Rounding based on Element-wise Division for Post-Training Quantization
di: Lee, Jung Hyun, et al.
Pubblicazione: (2023)

How Instruction and Reasoning Data shape Post-Training: Data Quality through the Lens of Layer-wise Gradients
di: Li, Ming, et al.
Pubblicazione: (2025)

Exploring Layer-wise Information Effectiveness for Post-Training Quantization in Small Language Models
di: Xiao, He, et al.
Pubblicazione: (2025)

Can You Share Your Story? Modeling Clients' Metacognition and Openness for LLM Therapist Evaluation
di: Kim, Minju, et al.
Pubblicazione: (2025)

From KMMLU-Redux to KMMLU-Pro: A Professional Korean Benchmark Suite for LLM Evaluation
di: Hong, Seokhee, et al.
Pubblicazione: (2025)

Motif-2-12.7B-Reasoning: A Practitioner's Guide to RL Training Recipes
di: Lim, Junghwan, et al.
Pubblicazione: (2025)

Two-Stage Grid Optimization for Group-wise Quantization of LLMs
di: Kim, Junhan, et al.
Pubblicazione: (2026)

The Impact of Post-training on Data Contamination
di: Kocyigit, Muhammed Yusuf, et al.
Pubblicazione: (2026)

Reasoning or Fluency? Dissecting Probabilistic Confidence in Best-of-N Selection
di: Kim, Hojin, et al.
Pubblicazione: (2026)

A Ridge Too Far: Correcting Over-Shrinkage via Negative Regularization
di: Kim, Dongseok, et al.
Pubblicazione: (2025)

SLIP & ETHICS: Graduated Intervention for AI Emotional Companions
di: Kim, Minseo
Pubblicazione: (2026)

Learning-Zone Energy: Online Data Selection for Efficient RL Post-Training
di: Cui, Peng, et al.
Pubblicazione: (2026)

CoScale-RL: Efficient Post-Training by Co-Scaling Data and Computation
di: Chen, Yutong, et al.
Pubblicazione: (2026)

InterPol: De-anonymizing LM Arena via Interpolated Preference Learning
di: Cho, Minsung, et al.
Pubblicazione: (2026)

Targeted Data Protection for Diffusion Model by Matching Training Trajectory
di: Lee, Hojun, et al.
Pubblicazione: (2025)

Pushing the Boundaries of Multiple Choice Evaluation to One Hundred Options
di: Lee, Nahyun, et al.
Pubblicazione: (2026)

Hybrid-LoRA: Bridging Full Fine-Tuning and Low-Rank Adaptation for Post-Training
di: Zhang, Chengqian, et al.
Pubblicazione: (2026)

Don't Look at the Numbers: Visual Anchoring Bias and Layer-wise Representation in VLMs
di: Shalankin, M.
Pubblicazione: (2026)

SPRInG: Continual LLM Personalization via Selective Parametric Adaptation and Retrieval-Interpolated Generation
di: Kim, Seoyeon, et al.
Pubblicazione: (2026)

Agent^2 RL-Bench: Can LLM Agents Engineer Agentic RL Post-Training?
di: Chen, Wanyi, et al.
Pubblicazione: (2026)

EMCEE: Improving Multilingual Capability of LLMs via Bridging Knowledge and Reasoning with Extracted Synthetic Multilingual Context
di: Koo, Hamin, et al.
Pubblicazione: (2025)

Enhancing Instruction Following of LLMs via Activation Steering with Dynamic Rejection
di: Kang, Minjae, et al.
Pubblicazione: (2026)

OPSD Compresses What RLVR Teaches: A Post-RL Compaction Stage for Reasoning Models
di: Kim, Jaehoon, et al.
Pubblicazione: (2026)

Detecting Data Contamination from Reinforcement Learning Post-training for Large Language Models
di: Tao, Yongding, et al.
Pubblicazione: (2025)

Layer-wise QUBO-Based Training of CNN Classifiers for Quantum Annealing
di: Atallah, Mostafa, et al.
Pubblicazione: (2026)

Information-Theoretic Greedy Layer-wise Training for Traffic Sign Recognition
di: Lyu, Shuyan, et al.
Pubblicazione: (2025)

From SFT to RL: Demystifying the Post-Training Pipeline for LLM-based Vulnerability Detection
di: Li, Youpeng, et al.
Pubblicazione: (2026)