Copertina

Learning from Failures: Understanding LLM Alignment through Failure-Aware Inverse RL

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Patel, Nyal, Bou, Matthieu, Jagota, Arjun, Krishna, Satyapriya, Parbhoo, Sonali
Natura:	Preprint
Pubblicazione:	2025
Soggetti:	Machine Learning Computation and Language
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

The Alignment Auditor: A Bayesian Framework for Verifying and Refining LLM Objectives
di: Bou, Matthieu, et al.
Pubblicazione: (2025)

Insights from the Inverse: Reconstructing LLM Training Goals Through Inverse Reinforcement Learning
di: Joselowitz, Jared, et al.
Pubblicazione: (2024)

Solving the Inverse Alignment Problem for Efficient RLHF
di: Krishna, Shambhavi, et al.
Pubblicazione: (2024)

On Almost Surely Safe Alignment of Large Language Models at Inference-Time
di: Ji, Xiaotong, et al.
Pubblicazione: (2025)

Accurate Failure Prediction in Agents Does Not Imply Effective Failure Prevention
di: Vasudev, Rakshith, et al.
Pubblicazione: (2026)

Inverse Reinforcement Learning with Dynamic Reward Scaling for LLM Alignment
di: Cheng, Ruoxi, et al.
Pubblicazione: (2025)

Learning from Failures in Multi-Attempt Reinforcement Learning
di: Chung, Stephen, et al.
Pubblicazione: (2025)

Tree-Based Leakage Inspection and Control in Concept Bottleneck Models
di: Ragkousis, Angelos, et al.
Pubblicazione: (2024)

GLIDE-RL: Grounded Language Instruction through DEmonstration in RL
di: Kharyal, Chaitanya, et al.
Pubblicazione: (2024)

Bayesian Inverse Transition Learning: Learning Dynamics From Near-Optimal Trajectories
di: Benac, Leo, et al.
Pubblicazione: (2024)

In-Context Explainers: Harnessing LLMs for Explaining Black Box Models
di: Kroeger, Nicholas, et al.
Pubblicazione: (2023)

Failure Modes of Maximum Entropy RLHF
di: Çağatan, Ömer Veysel, et al.
Pubblicazione: (2025)

Why LLMs Fail at Causal Discovery and How Interventional Agents Escape
di: Roy, Amartya, et al.
Pubblicazione: (2026)

Improving RL Exploration for LLM Reasoning through Retrospective Replay
di: Dou, Shihan, et al.
Pubblicazione: (2025)

Failure Modes of LLMs for Causal Reasoning on Narratives
di: Yamin, Khurram, et al.
Pubblicazione: (2024)

Diagnosing Structural Failures in LLM-Based Evidence Extraction for Meta-Analysis
di: Tan, Zhiyin, et al.
Pubblicazione: (2026)

Importing Phantoms: Measuring LLM Package Hallucination Vulnerabilities
di: Krishna, Arjun, et al.
Pubblicazione: (2025)

Understanding Likelihood Over-optimisation in Direct Alignment Algorithms
di: Shi, Zhengyan, et al.
Pubblicazione: (2024)

CoG: Controllable Graph Reasoning via Relational Blueprints and Failure-Aware Refinement over Knowledge Graphs
di: Liu, Yuanxiang, et al.
Pubblicazione: (2026)

A Theoretical Understanding of Self-Correction through In-context Alignment
di: Wang, Yifei, et al.
Pubblicazione: (2024)

TreeRL: LLM Reinforcement Learning with On-Policy Tree Search
di: Hou, Zhenyu, et al.
Pubblicazione: (2025)

Understanding How CodeLLMs (Mis)Predict Types with Activation Steering
di: Lucchetti, Francesca, et al.
Pubblicazione: (2024)

REA-RL: Reflection-Aware Online Reinforcement Learning for Efficient Reasoning
di: Deng, Hexuan, et al.
Pubblicazione: (2025)

TACO-RL: Task Aware Prompt Compression Optimization with Reinforcement Learning
di: Shandilya, Shivam, et al.
Pubblicazione: (2024)

From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization
di: Zhou, Chenxi, et al.
Pubblicazione: (2026)

TemplateRL: Structured Template-Guided Reinforcement Learning for LLM Reasoning
di: Wu, Jinyang, et al.
Pubblicazione: (2025)

Large Language Model Reasoning Failures
di: Song, Peiyang, et al.
Pubblicazione: (2026)

Replay Failures as Successes: Sample-Efficient Reinforcement Learning for Instruction Following
di: Zhang, Kongcheng, et al.
Pubblicazione: (2025)

Efficient Detection of Intermittent Job Failures Using Few-Shot Learning
di: Aïdasso, Henri, et al.
Pubblicazione: (2025)

Understanding the Effects of Iterative Prompting on Truthfulness
di: Krishna, Satyapriya, et al.
Pubblicazione: (2024)

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

Multi-Task GRPO: Reliable LLM Reasoning Across Tasks
di: Ramesh, Shyam Sundhar, et al.
Pubblicazione: (2026)

DualDiffusion: A Speculative Decoding Strategy for Masked Diffusion Models
di: Goyal, Satyam, et al.
Pubblicazione: (2026)

Do regularization methods for shortcut mitigation work as intended?
di: Hong, Haoyang, et al.
Pubblicazione: (2025)

Is the Reversal Curse a Binding Problem? Uncovering Limitations of Transformers from a Basic Generalization Failure
di: Wang, Boshi, et al.
Pubblicazione: (2025)

Mirage or Method? How Model-Task Alignment Induces Divergent RL Conclusions
di: Wu, Haoze, et al.
Pubblicazione: (2025)

PluralLLM: Pluralistic Alignment in LLMs via Federated Learning
di: Srewa, Mahmoud, et al.
Pubblicazione: (2025)

FP8-RL: A Practical and Stable Low-Precision Stack for LLM Reinforcement Learning
di: Qiu, Zhaopeng, et al.
Pubblicazione: (2026)

ResRL: Boosting LLM Reasoning via Negative Sample Projection Residual Reinforcement Learning
di: Lin, Zihan, et al.
Pubblicazione: (2026)

Mass-Producing Failures of Multimodal Systems with Language Models
di: Tong, Shengbang, et al.
Pubblicazione: (2023)