Imagen de Portada

Test-time RL alignment exposes task familiarity artifacts in LLM benchmarks

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Wang, Kun, Heckel, Reinhard
Formato:	Preprint
Publicado:	2026
Materias:	Machine Learning
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Measuring Fingerprints of Web-filtered Text Datasets and Fingerprint Propagation Through Training
por: Mansour, Youssef, et al.
Publicado: (2024)

A Deep Learning Method for Simultaneous Denoising and Missing Wedge Reconstruction in Cryogenic Electron Tomography
por: Wiedemann, Simon, et al.
Publicado: (2023)

Asymmetric Prompt Weighting for Reinforcement Learning with Verifiable Rewards
por: Heckel, Reinhard, et al.
Publicado: (2026)

Robustness of Deep Learning for Accelerated MRI: Benefits of Diverse Training Data
por: Lin, Kang, et al.
Publicado: (2023)

Trace Reconstruction with Language Models
por: Weindel, Franziska, et al.
Publicado: (2025)

Transformer-Based Decoding in Concatenated Coding Schemes Under Synchronization Errors
por: Streit, Julian, et al.
Publicado: (2025)

Reducing the Representation Error of GAN Image Priors Using the Deep Decoder
por: Daniels, Mara, et al.
Publicado: (2020)

Resolution-Robust 3D MRI Reconstruction with 2D Diffusion Priors: Diverse-Resolution Training Outperforms Interpolation
por: Krainovic, Anselm, et al.
Publicado: (2024)

Automated Program Repair: Emerging trends pose and expose problems for benchmarks
por: Renzullo, Joseph, et al.
Publicado: (2024)

Deep Learning for Accelerated and Robust MRI Reconstruction: a Review
por: Heckel, Reinhard, et al.
Publicado: (2024)

Spectral alignment of stochastic gradient descent for high-dimensional classification tasks
por: Arous, Gerard Ben, et al.
Publicado: (2023)

Online Detection of LLM-Generated Texts via Sequential Hypothesis Testing by Betting
por: Chen, Can, et al.
Publicado: (2024)

Offline Multi-task Transfer RL with Representational Penalization
por: Bose, Avinandan, et al.
Publicado: (2024)

Resurrecting saturated LLM benchmarks with adversarial encoding
por: Ivanov, Igor, et al.
Publicado: (2025)

Inference time LLM alignment in single and multidomain preference spectrum
por: Shahriar, Sadat, et al.
Publicado: (2024)

JigsawRL: Assembling RL Pipelines for Efficient LLM Post-Training
por: Hu, Zhengding, et al.
Publicado: (2026)

DUMP: Automated Distribution-Level Curriculum Learning for RL-based LLM Post-training
por: Wang, Zhenting, et al.
Publicado: (2025)

Slug Mobile: Test-Bench for RL Testing
por: Morris, Jonathan Wellington, et al.
Publicado: (2024)

Harnessing Test-time Adaptation for NLU tasks Involving Dialects of English
por: Nguyen, Duke, et al.
Publicado: (2025)

Detecting Proxy Gaming in RL and LLM Alignment via Evaluator Stress Tests
por: Shihab, Ibne Farabi, et al.
Publicado: (2025)

When Does Multi-Agent RL Improve LLM Workflows? Workflow, Scale, and Policy-Sharing Tradeoffs
por: Zeng, Yifan, et al.
Publicado: (2026)

Can time series forecasting be automated? A benchmark and analysis
por: Sreedhara, Anvitha Thirthapura, et al.
Publicado: (2024)

Enhancing RL Safety with Counterfactual LLM Reasoning
por: Gross, Dennis, et al.
Publicado: (2024)

ETS: Energy-Guided Test-Time Scaling for Training-Free RL Alignment
por: Li, Xiuyu, et al.
Publicado: (2026)

Putting the Value Back in RL: Better Test-Time Scaling by Unifying LLM Reasoners With Verifiers
por: Sareen, Kusha, et al.
Publicado: (2025)

Critique to Verify: Accurate and Honest Test-Time Scaling with RL-Trained Verifiers
por: Yang, Zhicheng, et al.
Publicado: (2025)

Language models scale reliably with over-training and on downstream tasks
por: Gadre, Samir Yitzhak, et al.
Publicado: (2024)

TIMIT Speaker Profiling: A Comparison of Multi-task learning and Single-task learning Approaches
por: Wang, Rong, et al.
Publicado: (2024)

Better audio representations are more brain-like: linking model-brain alignment with performance in downstream auditory tasks
por: Pepino, Leonardo, et al.
Publicado: (2025)

FREQuency ATTribution: benchmarking frequency-based occlusion for time series data
por: Mercier, Dominique, et al.
Publicado: (2025)

On Entropy Control in LLM-RL Algorithms
por: Shen, Han
Publicado: (2025)

Token-Efficient RL for LLM Reasoning
por: Lee, Alan, et al.
Publicado: (2025)

On the Generalization Gap in LLM Planning: Tests and Verifier-Reward RL
por: Belcamino, Valerio, et al.
Publicado: (2026)

LLM-Guided Search for Deletion-Correcting Codes
por: Weindel, Franziska, et al.
Publicado: (2025)

Deployment-complete benchmarking
por: Mansouri, El Mustapha, et al.
Publicado: (2026)

Statistical benchmarking of transformer models in low signal-to-noise time-series forecasting
por: Garcia, Cyril, et al.
Publicado: (2026)

SPA-RL: Reinforcing LLM Agents via Stepwise Progress Attribution
por: Wang, Hanlin, et al.
Publicado: (2025)

Optimistic Critic Reconstruction and Constrained Fine-Tuning for General Offline-to-Online RL
por: Luo, Qin-Wen, et al.
Publicado: (2024)

Learning to Trust Bellman Updates: Selective State-Adaptive Regularization for Offline RL
por: Luo, Qin-Wen, et al.
Publicado: (2025)

FuzzingRL: Reinforcement Fuzz-Testing for Revealing VLM Failures
por: Xu, Jiajun, et al.
Publicado: (2026)