RubricEval: A Rubric-Level Meta-Evaluation Benchmark for LLM Judges in Instruction Following

Rubric-based evaluation has become a prevailing paradigm for evaluating instruction following in large language models (LLMs). Despite its widespread use, the reliability of these rubric-level evaluations remains unclear, calling for meta-evaluation. However, prior meta-evaluation efforts largely focus on the response level, failing to assess the fine-grained judgment accuracy that rubric-based evaluation relies on. To bridge this gap, we introduce RubricEval. Our benchmark features: (1) the first rubric-level meta-evaluation benchmark for instruction following, (2) diverse instructions and responses spanning multiple categories and model sources, and (3) a substantial set of 3,486 quality-controlled instances, along with Easy/Hard subsets that better differentiates judge performance. Our experiments reveal that rubric-level judging remains far from solved: even GPT-4o, a widely adopted judge in instruction-following benchmarks, achieves only 55.97% on Hard subset. Considering evaluation paradigm, rubric-level evaluation outperforms checklist-level, explicit reasoning improves accuracy, and both together reduce inter-judge variance. Through our established rubric taxonomy, we further identify common failure modes and offer actionable insights for reliable instruction-following evaluation.

From AI Assistant to AI Scientist: Autonomous Discovery of LLM-RL Algorithms with LLM Agents

Discovering improved policy optimization algorithms for language models remains a costly manual process requiring repeated mechanism-level modification and validation. Unlike simple combinatorial code search, this problem requires searching over algorithmic mechanisms tightly coupled with training dynamics while reusing empirical evidence across iterations. We propose POISE, a closed-loop framework for automated discovery of policy optimization algorithms for language models. POISE maintains a structured, genealogically linked archive linking proposals, executable implementations, standardized evaluations, and natural-language reflections to support evidence-driven iteration. In mathematical reasoning experiments starting from GRPO, POISE evaluates 64 candidate algorithms and discovers improved mechanisms, including analytic-variance scaling and validity masking. The best variant improves weighted Overall from 47.8 to 52.5 (+4.6) and increases AIME25 pass@32 from 26.7% to 43.3%, demonstrating the feasibility of automated policy optimization discovery while supporting interpretable design principles.

Thinking with Constructions: A Benchmark and Policy Optimization for Visual-Text Interleaved Geometric Reasoning

Geometric reasoning inherently requires "thinking with constructions" -- the dynamic manipulation of visual aids to bridge the gap between problem conditions and solutions. However, existing Multimodal Large Language Models (MLLMs) are largely confined to passive inference with static diagrams, lacking the strategic knowledge of when and how to construct effective visual aids. To address this, we present a framework for Visual-Text Interleaved Chain-of-Thought. We first introduce GeoAux-Bench, the first benchmark comprising 4,334 geometry problems that aligns textual construction steps with ground-truth visual updates. Our pilot study reveals two critical insights: (1) interleaved visual-textual aids outperform single-modality counterparts, which cannot losslessly capture geometric synergy; and (2) valid constructions act as entropy reducers, strongly correlating with reduced reasoning perplexity. Building on these findings, we propose Action Applicability Policy Optimization (A2PO), a reinforcement learning paradigm for mastering strategic construction. A2PO employs Adaptive Reward Shaping to regulate the timing and quality of visual aids via counterfactual sampling to distinguish necessary from redundant constructions. Experiments demonstrate our approach enables MLLMs to leverage selective auxiliary constructions, yielding a 3.51% gain over strong baselines. Code and data are available on GitHub.

DIVE: Scaling Diversity in Agentic Task Synthesis for Generalizable Tool Use

Recent work synthesizes agentic tasks for post-training tool-using LLMs, yet robust generalization under shifts in tasks and toolsets remains an open challenge. We trace this brittleness to insufficient diversity in synthesized tasks. Scaling diversity is difficult because training requires tasks to remain executable and verifiable, while generalization demands coverage of diverse tool types, toolset combinations, and heterogeneous tool-use patterns. We propose DIVE, an evidence-driven recipe that inverts synthesis order, executing diverse, real-world tools first and reverse-deriving tasks strictly entailed by the resulting traces, thereby providing grounding by construction. DIVE scales structural diversity along two controllable axes, tool-pool coverage and per-task toolset variety, and an Evidence Collection--Task Derivation loop further induces rich multi-step tool-use patterns across 373 tools in five domains. Training Qwen3-8B on DIVE data (48k SFT + 3.2k RL) improves by +22 average points across 9 OOD benchmarks and outperforms the strongest 8B baseline by +68. Remarkably, controlled scaling analysis reveals that diversity scaling consistently outperforms quantity scaling for OOD generalization, even with 4x less data.

CODA: Difficulty-Aware Compute Allocation for Adaptive Reasoning

The emergence of large reasoning models demonstrates that scaling inference-time compute significantly enhances performance on complex tasks. However, it often falls into another trap: overthinking simple problems, where repetitive rationales yield minimal accuracy gains at a disproportionately high cost. This motivates adaptive reasoning: dynamically aligning reasoning depth with instance difficulty. In this paper, we study adaptive reasoning from an optimality perspective, formalizing it as a utility maximization problem where tokens are allocated until the marginal accuracy gain falls below the incremental cost. Based on this, we propose CODA (Compute Allocation by Difficulty Awareness), a method that operationalizes this principle by allocating tokens via a policy-internal difficulty signal. Specifically, CODA estimates difficulty via group-based rollouts and maps it to two non-negative gates that modulate a length-dependent shaping term on top of the binary base reward. The easy-side gate penalizes verbosity on simple instances, whereas the hard-side gate encourages more deliberative rollouts on challenging ones. Across model scales and benchmarks, CODA achieves adaptive reasoning without external annotations or user-provided budgets: on easy tasks, CODA reduces token costs by over 60% while maintaining strong accuracy, whereas on hard tasks it incentivizes more deliberative rollouts to maximize performance.

HER: Human-like Reasoning and Reinforcement Learning for LLM Role-playing

LLM role-playing, i.e., using LLMs to simulate specific personas, has emerged as a key capability in various applications, such as companionship, content creation, and digital games. While current models effectively capture character tones and knowledge, simulating the inner thoughts behind their behaviors remains a challenge. Towards cognitive simulation in LLM role-play, previous efforts mainly suffer from two deficiencies: data with high-quality reasoning traces, and reliable reward signals aligned with human preferences. In this paper, we propose HER, a unified framework for cognitive-level persona simulation. HER introduces dual-layer thinking, which distinguishes characters' first-person thinking from LLMs' third-person thinking. To bridge these gaps, we curate reasoning-augmented role-playing data via reverse engineering and construct human-aligned principles and reward models. Leveraging these resources, we train \method models based on Qwen3-32B via supervised and reinforcement learning. Extensive experiments validate the effectiveness of our approach. Notably, our models significantly outperform the Qwen3-32B baseline, achieving a 30.26 improvement on the CoSER benchmark and a 14.97 gain on the Minimax Role-Play Bench. Our datasets, principles, and models will be released to facilitate future research.

HUMANLLM: Benchmarking and Reinforcing LLM Anthropomorphism via Human Cognitive Patterns

Large Language Models (LLMs) have demonstrated remarkable capabilities in reasoning and generation, serving as the foundation for advanced persona simulation and Role-Playing Language Agents (RPLAs). However, achieving authentic alignment with human cognitive and behavioral patterns remains a critical challenge for these agents. We present HUMANLLM, a framework treating psychological patterns as interacting causal forces. We construct 244 patterns from ~12,000 academic papers and synthesize 11,359 scenarios where 2-5 patterns reinforce, conflict, or modulate each other, with multi-turn conversations expressing inner thoughts, actions, and dialogue. Our dual-level checklists evaluate both individual pattern fidelity and emergent multi-pattern dynamics, achieving strong human alignment (r=0.91) while revealing that holistic metrics conflate simulation accuracy with social desirability. HUMANLLM-8B outperforms Qwen3-32B on multi-pattern dynamics despite 4x fewer parameters, demonstrating that authentic anthropomorphism requires cognitive modeling--simulating not just what humans do, but the psychological processes generating those behaviors.

Outcome-Grounded Advantage Reshaping for Fine-Grained Credit Assignment in Mathematical Reasoning

Group Relative Policy Optimization (GRPO) has emerged as a promising critic-free reinforcement learning paradigm for reasoning tasks. However, standard GRPO employs a coarse-grained credit assignment mechanism that propagates group-level rewards uniformly to to every token in a sequence, neglecting the varying contribution of individual reasoning steps. We address this limitation by introducing Outcome-grounded Advantage Reshaping (OAR), a fine-grained credit assignment mechanism that redistributes advantages based on how much each token influences the model's final answer. We instantiate OAR via two complementary strategies: (1) OAR-P, which estimates outcome sensitivity through counterfactual token perturbations, serving as a high-fidelity attribution signal; (2) OAR-G, which uses an input-gradient sensitivity proxy to approximate the influence signal with a single backward pass. These importance signals are integrated with a conservative Bi-Level advantage reshaping scheme that suppresses low-impact tokens and boosts pivotal ones while preserving the overall advantage mass. Empirical results on extensive mathematical reasoning benchmarks demonstrate that while OAR-P sets the performance upper bound, OAR-G achieves comparable gains with negligible computational overhead, both significantly outperforming a strong GRPO baseline, pushing the boundaries of critic-free LLM reasoning.

Structured Reasoning for Large Language Models

Large language models (LLMs) achieve strong performance by generating long chains of thought, but longer traces always introduce redundant or ineffective reasoning steps. One typical behavior is that they often perform unnecessary verification and revisions even if they have reached the correct answers. This limitation stems from the unstructured nature of reasoning trajectories and the lack of targeted supervision for critical reasoning abilities. To address this, we propose Structured Reasoning (SCR), a framework that decouples reasoning trajectories into explicit, evaluable, and trainable components. We mainly implement SCR using a Generate-Verify-Revise paradigm. Specifically, we construct structured training data and apply Dynamic Termination Supervision to guide the model in deciding when to terminate reasoning. To avoid interference between learning signals for different reasoning abilities, we adopt a progressive two-stage reinforcement learning strategy: the first stage targets initial generation and self-verification, and the second stage focuses on revision. Extensive experiments on three backbone models show that SCR substantially improves reasoning efficiency and self-verification. Besides, compared with existing reasoning paradigms, it reduces output token length by up to 50%.

LSRIF: Logic-Structured Reinforcement Learning for Instruction Following

Instruction-following is critical for large language models, but real-world instructions often contain logical structures such as sequential dependencies and conditional branching. Existing methods typically construct datasets with parallel constraints and optimize average rewards, ignoring logical dependencies and yielding noisy signals. We propose a logic-structured training framework LSRIF that explicitly models instruction logic. We first construct a dataset LSRInstruct with constraint structures such as parallel, sequential, and conditional types, and then design structure-aware rewarding method LSRIF including average aggregation for parallel structures, failure-penalty propagation for sequential structures, and selective rewards for conditional branches. Experiments show LSRIF brings significant improvements in instruction-following (in-domain and out-of-domain) and general reasoning. Analysis reveals that learning with explicit logic structures brings parameter updates in attention layers and sharpens token-level attention to constraints and logical operators.