SARL rewards reasoning topology to improve label-free RL, outperforming baselines with gains up to 44.7% on math and 34.6% on open-ended tasks while maintaining more stable training.
hub Mixed citations
arXiv preprint arXiv:2510.07743 , year=
Mixed citation behavior. Most common role is background (60%).
22
Pith papers citing it
Background
60% of classified citations
hub tools
citation-role summary
background 3
dataset 1
method 1
citation-polarity summary
years
2026 22representative citing papers
RGSD distills rubric-conditioned teacher distributions into base policies token-by-token, matching GRPO rubric satisfaction on Qwen models with one rollout and zero verifier calls.
SVR learns a bank of contrastive rubrics from preference data via max-margin boundaries and prompt-conditioned selection, narrowing the gap to human rubrics on RubricBench from 24.1 to 0.3 points.
Co-ReAct adds step-level rubric guidance to ReAct agents via a GRPO-trained generator using list-wise ranking rewards, yielding consistent gains on DeepResearchBench and SQA-CS-V2.
RDPO applies magnitude-aware quantile normalization and Mahalanobis whitening to decorrelate heterogeneous rewards in multi-objective RL, improving instruction following and writing quality on LongCat-Flash post-training while staying competitive on reasoning and coding.
Think-with-Rubrics has LLMs generate rubrics internally before responding, outperforming external rubric-as-reward baselines by 3.87 points on average across benchmarks.
Rubric-based on-policy distillation allows training student models using only teacher responses by generating scoring rubrics from contrasts and using them for on-policy optimization, achieving superior performance and up to 10x better sample efficiency than logit-based approaches.
rDPO uses offline-built rubrics to generate on-policy preference data for DPO, raising benchmark scores in visual tasks over outcome-based filtering and style baselines.
ARCO introduces a co-evolving rubric model with generation and scoring heads plus a trajectory decomposition constraint that improves exact-match scores on multi-hop QA tasks over outcome, rubric, and process reward baselines.
Skill-RM unifies heterogeneous reward criteria by modeling reward computation as dynamic execution of a reusable Reward-Evaluation Skill within an agent framework.
DR-rubric is a two-stage framework using iterative agentic search to generate atomic verifiable constraints for GRPO-based RL, achieving competitive performance on 6 benchmarks with 1K-3K examples via bootstrap or frontier-model rubrics.
PARL formulates personalized LLM evaluation as a learning problem that induces preference-aware rubrics from raw user histories via discriminative RL and self-validation.
RUBRIC-ARROW is an alternating rubric generator and judge framework that uses probability-based scoring and pairwise preferences to improve pointwise reward modeling accuracy for LLM post-training in non-verifiable domains.
MERIT trains a small reviewer assessor via rubric-guided RL with LLM rewards and distills it to a SOTA embedding retriever for paper-reviewer matching on LR-Bench and CMU Gold.
GrowLoop proposes a human-seeded self-evolving framework that co-evolves rubrics and cases to evaluate conversational human-likeness with differentiated agreement rules.
POW3R adapts rubric criterion weights via rollout contrast in RLVR to improve mean reward, strict completion rates, and training speed over static rubric aggregation on multimodal and text tasks.
DeltaRubric decomposes multimodal preference evaluation into self-generated planning and verification steps within a single model, producing large accuracy improvements on VL-RewardBench via multi-role reinforcement learning.
Five expert-authored clinical scenarios with atomic weighted rubrics show frontier LLMs passing only 32-42% of critical criteria versus 80-90% of low-stakes ones, with 52% of critical criteria failed by all models tested.
A new pipeline uses interpretability to characterize concepts in preference data and shape rewards via feature or data interventions during LM post-training.
RLR³ extends RLVR to criterion-level rubric verification via dual execution paths, minimal exposure masking, hierarchical aggregation, and saturation mitigation, delivering 4.7-point gains over base on 15 benchmarks with Qwen3-VL-30B-A3B.
The paper introduces the Proxy Compression Hypothesis as a unifying framework explaining reward hacking in RLHF as an emergent result of compressing high-dimensional human objectives into proxy reward signals under optimization pressure.
Proxy RL produces a staged proxy-internalization capability that emerges before and predicts reward hacking in coding environments.
citing papers explorer
-
Rubric-Guided Self-Distillation: Post-Training Without Rubric Verifiers
RGSD distills rubric-conditioned teacher distributions into base policies token-by-token, matching GRPO rubric satisfaction on Qwen models with one rollout and zero verifier calls.
-
Support Vector Rubrics: Closing the Gap Between Self-Generated and Human Rubrics
SVR learns a bank of contrastive rubrics from preference data via max-margin boundaries and prompt-conditioned selection, narrowing the gap to human rubrics on RubricBench from 24.1 to 0.3 points.
-
Co-ReAct: Rubrics as Step-Level Collaborators for ReAct Agents
Co-ReAct adds step-level rubric guidance to ReAct agents via a GRPO-trained generator using list-wise ranking rewards, yielding consistent gains on DeepResearchBench and SQA-CS-V2.
-
Multi-Objective and Mixed-Reward Reinforcement Learning via Reward-Decorrelated Policy Optimization
RDPO applies magnitude-aware quantile normalization and Mahalanobis whitening to decorrelate heterogeneous rewards in multi-objective RL, improving instruction following and writing quality on LongCat-Flash post-training while staying competitive on reasoning and coding.
-
Think-with-Rubrics: From External Evaluator to Internal Reasoning Guidance
Think-with-Rubrics has LLMs generate rubrics internally before responding, outperforming external rubric-as-reward baselines by 3.87 points on average across benchmarks.
-
Rubric-based On-policy Distillation
Rubric-based on-policy distillation allows training student models using only teacher responses by generating scoring rubrics from contrasts and using them for on-policy optimization, achieving superior performance and up to 10x better sample efficiency than logit-based approaches.
-
Visual Preference Optimization with Rubric Rewards
rDPO uses offline-built rubrics to generate on-policy preference data for DPO, raising benchmark scores in visual tasks over outcome-based filtering and style baselines.
-
ARCO: Adaptive Rubric with Co-Evolution for Multi-Step LLM-Based Agents
ARCO introduces a co-evolving rubric model with generation and scoring heads plus a trajectory decomposition constraint that improves exact-match scores on multi-hop QA tasks over outcome, rubric, and process reward baselines.
-
Skill-RM: Unifying Heterogeneous Evaluation Criteria via Agent Skill
Skill-RM unifies heterogeneous reward criteria by modeling reward computation as dynamic execution of a reusable Reward-Evaluation Skill within an agent framework.
-
Deep Research as Rubric for Reinforcement Learning
DR-rubric is a two-stage framework using iterative agentic search to generate atomic verifiable constraints for GRPO-based RL, achieving competitive performance on 6 benchmarks with 1K-3K examples via bootstrap or frontier-model rubrics.
-
Preference-Aware Rubric Learning for Personalized Evaluation
PARL formulates personalized LLM evaluation as a learning problem that induces preference-aware rubrics from raw user histories via discriminative RL and self-validation.
-
RUBRIC-ARROW: Alternating Pointwise Rubric Reward Modeling for LLM Post-training in Non-verifiable Domains
RUBRIC-ARROW is an alternating rubric generator and judge framework that uses probability-based scoring and pairwise preferences to improve pointwise reward modeling accuracy for LLM post-training in non-verifiable domains.
-
MERIT: Matching Expertise via Rubric-Informed Training for Reviewer Assignment
MERIT trains a small reviewer assessor via rubric-guided RL with LLM rewards and distills it to a SOTA embedding retriever for paper-reviewer matching on LR-Bench and CMU Gold.
-
GrowLoop: Self-Evolving Conversation Evaluation Seeded by Human
GrowLoop proposes a human-seeded self-evolving framework that co-evolves rubrics and cases to evaluate conversational human-likeness with differentiated agreement rules.
-
Not Every Rubric Teaches Equally: Policy-Aware Rubric Rewards for RLVR
POW3R adapts rubric criterion weights via rollout contrast in RLVR to improve mean reward, strict completion rates, and training speed over static rubric aggregation on multimodal and text tasks.
-
DeltaRubric: Generative Multimodal Reward Modeling via Joint Planning and Verification
DeltaRubric decomposes multimodal preference evaluation into self-generated planning and verification steps within a single model, producing large accuracy improvements on VL-RewardBench via multi-role reinforcement learning.
-
A rubric-based controlled comparison of frontier language models on expert-authored clinical reasoning tasks
Five expert-authored clinical scenarios with atomic weighted rubrics show frontier LLMs passing only 32-42% of critical criteria versus 80-90% of low-stakes ones, with 52% of critical criteria failed by all models tested.
-
Anatomy of Post-Training: Using Interpretability to Characterize Data and Shape the Learning Signal
A new pipeline uses interpretability to characterize concepts in preference data and shape rewards via feature or data interventions during LM post-training.
-
Reinforcement Learning with Robust Rubric Rewards
RLR³ extends RLVR to criterion-level rubric verification via dual execution paths, minimal exposure masking, hierarchical aggregation, and saturation mitigation, delivering 4.7-point gains over base on 15 benchmarks with Qwen3-VL-30B-A3B.
-
Reward Hacking in the Era of Large Models: Mechanisms, Emergent Misalignment, Challenges
The paper introduces the Proxy Compression Hypothesis as a unifying framework explaining reward hacking in RLHF as an emergent result of compressing high-dimensional human objectives into proxy reward signals under optimization pressure.
-
Proxy Reward Internalization and Mechanistic Exploitation: A Learned Precursor to Reward Hacking and Its Generalization
Proxy RL produces a staged proxy-internalization capability that emerges before and predicts reward hacking in coding environments.