Reliable question answering requires identifying not only whether an answer is correct, but also which available knowledge the prediction depends on. In realistic LLM-based QA, this knowledge may come from context, retrieval, decomposition, or intermediate reasoning, forming a noisy and redundant candidate space rather than a clean gold evidence set. We study \emph{knowledge dependency estimation}: estimating the sensitivity of a fixed black-box QA model to different candidate knowledge units. The challenge is to obtain fine-grained dependency scores without exhaustive test-time perturbation while modeling redundancy, substitutability, and complementarity. We propose \textbf{Knot}, a structured rank-aware knowledge dependency estimator. Knot learns from subset-level counterfactual supervision, models subset sensitivity through coverage over latent dependency factors, and derives rank-aware unit scores to identify influential candidates. Across multiple-choice and generative QA benchmarks, Knot outperforms all compared baselines in subset-sensitivity prediction and produces more faithful unit rankings than deployable baselines without extra QA-model calls; when used for practical risk screening, its dependency scores help flag error-prone QA predictions early.

Existing agent memory systems universally follow what we term a Memory-as-Tool paradigm where a single query triggers one-shot retrieval of flat passage lists, suffering from passive invocation, reasoning-retrieval decoupling, and structural mismatch between retrieved fragments and the agent's navigational needs. We propose MemCog, a Memory-as-Cognition system that makes memory access an integral part of the reasoning process. MemCog organizes user knowledge as Navigable Memory Store with associative link graphs, exposes Cross-Dimensional Navigation Interface for multi-step reasoning-driven traversal, and employs Proactive Reasoning Protocol that drives agents to spontaneously initiate memory exploration from conversational context. We additionally construct ProactiveMemBench, the first benchmark for evaluating proactive memory triggering. Experiments show that MemCog achieves state-of-the-art on passive QA benchmarks (92.98 on LoCoMo, 95.8 on LongMemEval) while substantially outperforming baselines on ProactiveMemBench, demonstrating the advantage of Memory-as-Cognition.

Cited RAG evaluation often treats visible sources as a grounding signal, but a real, topically relevant citation can still under-warrant the attached wording. We study this diagnostic failure as citation laundering: a related source is presented as warrant for an over-strong claim. We introduce FORCEBENCH, a contrastive stress test for evidence-force calibration. Each item holds a cited passage fixed and pairs an evidence-calibrated claim with a localized force-raised variant across five operational axes: relation, modality, scope, temporal validity, and numeric specificity. A calibrated evaluator should score the evidence-calibrated claim higher. Headline experiments use a fixed, locality-filtered 198-pair evaluation set. A citation-presence sanity check is uninformative by design; token and entity overlap still violate monotonicity on 32.8--36.4% of pairs. Across four reported model judges, standard generic support prompting is insufficient for this force-calibration stress test (aggregate MVR 47.2%), while explicit warrant-strength prompting lowers MVR to 24.5% but remains imperfect. We release the benchmark, prompts, outputs, and plug-in pipeline so citation evaluators can report monotonicity violation rate and force sensitivity alongside conventional support metrics.

Modern large language models (LLMs) achieve state-of-the-art machine translation performance, but they do so as broad generalists largely trained for many tasks and capabilities unrelated to translation. Thus, they are heavily overparameterized for this task, resulting in excessive memory and compute requirements. In this paper, we present a method for aggressively pruning experts from modern mixture-of-experts LLMs while incurring negligible degradation in translation quality. Our approach exploits expert specialization and the separability of multilingual capabilities in LLMs to identify experts irrelevant to translation. And because of the modular nature of MoEs, these can be easily pruned without any training. Without retraining, we are able to prune half of all experts with negligible degradation and 70% with only minor losses. With a very short SFT, we prune 75% of experts while recovering baseline performance, and in some settings remove nearly 90% while maintaining reasonable translation quality. Overall, our results show that translation requires only a fraction of the LLM, enabling substantial compression of the MoE blocks that contain over 90% of parameters.

Prompt-based personality control is a key technique for designing large language model (LLM) dialogue agents that behave consistently across social contexts. However, specifying Big Five personality traits (BFTs) in a prompt does not ensure that the intended traits are expressed in generated utterances. This paper investigates this mismatch from an interactionist perspective, viewing personality expression as a context-dependent outcome shaped by the interplay between trait specification and situational factors. We analyze how perceived BFT expression in LLM-generated dialogue is influenced by three prompt factors: personality traits, dialogue roles, and expressive styles. Using a factorial design that combines six personality conditions, three roles, and three expressive-style conditions, we generate 1,080 LLM-agent dialogues in each of English and Japanese. We then evaluate the target agent's utterances using an LLM-as-a-judge framework to estimate expressed Big Five traits. The results show that expressed personality is shaped not only by explicit trait specification, but also by dialogue role and expressive style. These effects are trait-specific: dialogue role strongly influences Openness, expressive style substantially shapes Conscientiousness and Agreeableness, and explicit trait specification dominates Neuroticism. Even without explicit personality-trait specification, social and expressive conditions induce distinct personality-like impressions. Cross-linguistic comparisons show broadly similar patterns between English and Japanese dialogues, with noticeable differences only under specific combinations of personality, role, and expressive style. These findings suggest that personality control in LLM agents should be understood not as a direct consequence of trait prompting, but as a context-dependent process involving personality specification, social role, and expressive style.

Diffusion models steer conditional generation with a tunable guidance scale to trade off prompt alignment and diversity. However, existing debiasing techniques are optimized for a single scale, degrading fairness when users adjust this parameter. We trace this behavior to a previously overlooked source by decomposing total bias into two components: a model bias and a guidance bias. While prior work primarily targets the former, we show that the guidance bias grows monotonically with the guidance scale, eventually dominating the high-guidance regimes users prefer. To address this, we extend Strong Demographic Parity to guidance and derive a condition under which the target distribution retains its group ratio across guidance scales. We propose StayFair, which leverages this condition to design fair guidance algorithms in both regimes. For classifier guidance, it equalizes the classifier's output distributions across groups; for classifier-free guidance, it shifts the null embedding by a prompt-dependent offset. Because StayFair modifies only the guidance step, it is orthogonal to model debiasing and can be layered onto existing fair diffusion models to extend their fairness across guidance scales. Across class-conditional and text-to-image generation, StayFair decouples fairness from the guidance scale without sacrificing image quality.

In recent years, Multi-Talker Audio-Video Generation (MTAVG) models have shown promising performance on fundamental metrics such as lip-sync and audio-visual alignment. However, these metrics remain insufficient for assessing cinematic expressiveness in scene-level generation. In multi-character scenes, generation models must go beyond audio-visual realism to convey coherent character performance and other higher-level cinematic qualities. To fill this gap, we introduce MTAVG-Bench 2.0, a benchmark for diagnosing failure modes of cinematic expressiveness in multi-talker audio-video generation. Unlike prior settings that mainly focus on the quality of basic multi-turn dialogue, MTAVG-Bench 2.0 targets short-drama and scene-level generation, and establishes a high-level failure taxonomy spanning acting, narrative, atmosphere, and audio-visual language. Based on this taxonomy, we construct more than 10,000 question-answering evaluation instances, together with subsets for short-drama-level assessment and temporal localization of failure modes, to systematically evaluate the ability of omni large language models to diagnose high-level audio-visual failures. Experimental results show that commercial omni models such as Gemini substantially outperform other evaluators, yet even the strongest models continue to struggle with complex failures in our benchmark. These results demonstrate that MTAVG-Bench 2.0 provides a systematic benchmark for failure diagnosis in cinematic multi-talker audio-video generation.

Clark Hash is a small method for storing neural embeddings in less space. It normalizes each database vector, applies a deterministic sparse signed Johnson-Lindenstrauss projection, clips the result, and stores a fixed-width scalar-quantized code. Queries stay in floating point and are scored against the stored sketches. In the default 384-dimensional sentence-embedding setting, Clark Hash stores a cosine-search vector in 48 bytes instead of 1536 bytes for dense f32 storage. This is 32x smaller. The method does not need a training pass, learned codebooks, rotations, or corpus statistics before new vectors can be stored. We describe the codec, the Rust implementation, and a multilingual sentence-similarity evaluation on 9,304 labeled pairs from 29 subsets. With a multilingual MiniLM encoder, the 48-byte sketches reached 0.910 and 0.946 macro Pearson correlation with dense cosine scores on STS17 and STS22. Clark Hash is not a new Johnson-Lindenstrauss theorem and it is not a replacement for approximate nearest-neighbor indexes. It is a simple stateless codec for compact embedding storage.

Instructing robots from demonstrations can be done through different teaching modalities, each with different usability and performance trade-offs. This paper compares kinesthetic guidance, joystick teleoperation, and hand gestures in a user study with eight participants. We evaluate replay success, modified NASA-TLX workload, and common teaching errors across three manipulation tasks. Kinesthetic guidance produced the shortest demonstrations, lowest workload, and highest success on the more orientation-sensitive and contact-rich tasks. Joystick teleoperation performed best on simple peg picking. Hand-gesture teaching, although less reliable overall, performed better than expected and in some cases achieved results comparable to kinesthetic guidance.

Large Language Models are increasingly applied in the petroleum industry, highlighting the need for a domain-specific evaluation framework. This study develops a benchmark for LLMs in petroleum engineering, including a three-stage process of data preprocessing, quality filtering, and multi-model validation. Using expert review, a standardized question bank with strong domain relevance and discriminative capability was constructed. The benchmark covers production, reservoir, and drilling engineering, with 1,200 questions across multiple-choice, true or false, term definition, and short-answer formats. Eight mainstream LLMs were evaluated under a unified API environment. Results show that models performed better on subjective than objective questions, indicating weaknesses in factual knowledge discrimination. The highest accuracies for multiple-choice and true or false questions were 65.3% and 74.3%, respectively. Gemini-3-Pro, Kimi-K2.5, and Claude-Opus-4.6-Thinking achieved the best overall scores of 72%-74%. Models performed best in production engineering and weakest in reservoir engineering. Chinese models showed advantages in multiple-choice questions, while international models performed slightly better in short-answer questions. The benchmark provides a reproducible and practical reference for evaluating and deploying LLMs in petroleum engineering.

Fine-tuning large language models often undermines their safety alignment, a problem further amplified by harmful fine-tuning attacks in which adversarial data removes safeguards and induces unsafe behaviors. We propose SPARD, a defense framework that integrates Safety-Projected Alternating optimization with Relevance-Diversity aware data selection. SPARD employs SPAG, which optimizes alternatively between utility updates and explicit safety projections with a set of safe data to enforce safety constraints. To curate safe data, we introduce a Relevance-Diversity Determinantal Point Process to select compact safe data, balancing task relevance and safety coverage. Experiments on GSM8K and OpenBookQA under four harmful fine-tuning attacks demonstrate that SPARD consistently achieves the lowest average attack success rates, substantially outperforming state-of-the-art defense methods, while maintaining high task accuracy. Code is available at https://github.com/shuhao02/SPARD.

Group Relative Policy Optimization (GRPO) is widely used for training reasoning models, but updating all sampled completions in each group incurs substantial cost and can reinforce verbose reasoning trajectories. In this paper, we study whether all completions provide equally useful update signals in GRPO-style reasoning RL. Our gradient-similarity analysis shows that, within the same prompt group, same-class completions often induce highly similar update directions, whereas correct-incorrect pairs provide more distinct contrastive signals. Motivated by this observation, we propose Binary Prefix Policy Optimization (BPPO), which uses the shortest correct completion and the shortest incorrect completion as a compact update unit while preserving full-group advantage normalization. BPPO further improves efficiency with adaptive completion scheduling and prefix-focused optimization; by updating only response prefixes, it avoids reinforcing redundant suffixes and encourages more concise responses. Experiments on GSM8K, MATH, and Geo3K show that BPPO achieves up to 6.08x speedup over GRPO while maintaining competitive accuracy, and reduces mean response length by approximately 30-50% without modifying the reward with an explicit length penalty.

Large language models (LLMs) are increasingly used to provide public-facing health information, yet existing safety evaluations overlook whether responses preserve comparable medical information across different user phrasings of the same question. To address this, we introduce the Medical Information Response Audit (MIRA), a bilingual, controlled benchmark that assesses whether LLMs provide comparable medical information across user-side language, register, and health literacy signals. MIRA contains 4,320 prompts built from 60 medically reviewed, low-risk health questions. Across five mainstream LLMs, models answered all medical questions, but responses to low health-literacy signals consistently omitted more key information, provided fewer concrete next steps, and offered less support for independent judgment. We term this pattern Differential Information Dilution (DID). Language effects are model-specific rather than uniformly worse for non-English prompts. A comparison with 300 real-world health queries provides preliminary evidence of rank-order validity. A knowledge-guided mitigation prompt reduces information dilution for most models, with the largest reductions in underinformative simplification observed for Claude (~8%) and Qwen (~6%).

Visual captioning requires models to capture visual content faithfully while minimizing both omission and hallucination. As the dominant paradigm for captioning, MLLMs have achieved strong performance through scaling and high-quality data. Recently, RL has emerged as a key route to driving MLLMs toward higher precision and broader coverage, however, existing reward designs for captioning fail to provide fine-grained and reliable signals for factual verification, limiting their effectiveness. To address this, we propose VCap, a Witness-Adjudicator reward that pairs the reference caption (a witness) with the visual signal (an adjudicator). By explicitly verifying factual consistency between the reference and policy-generated captions grounded in the visual signal, VCap delivers a reward signal with hypergeometric-distribution-level precision for caption quality verification. This design enables effective learning even from imperfect references, facilitating weak-to-strong generalization in RL training. In our experiments, an 8B model trained with VCap outperforms open- and closed-source SOTA models on multiple image and video captioning benchmarks. Human evaluation further confirms its strong alignment with factual correctness. Additionally, VCap improves MLLM perceptual capability, generalizes across tasks, and surpasses best-of-N distillation, challenging prior assumptions about RLVR.

LLMs for code generation are commonly evaluated in repeated-sampling settings using Pass@k, where multiple candidate programs are executed against unit tests under a finite sampling budget. While recent verifier-based reinforcement learning (RLVR) methods improve executable correctness, how these objectives affect redundancy among sampled programs remains poorly understood. In this work, we study implementation-level redundancy in code generation using JPlag, a plagiarism-detection system for code. Across models and benchmarks, we show that correctness-only RLVR often concentrates generations around repeated implementations, whereas Pass@k-aware objectives maintain lower redundancy and improve larger-budget performance. Motivated by these observations, we augment RLVR with direct anti-redundancy rewards based on JPlag similarity. Across 3 models and 3 benchmarks, discouraging near-duplicate generations reliably improves finite-budget executable performance, often matching or outperforming specialized Pass@k-aware objectives.

Out-of-distribution (OOD) detection is essential for deploying machine learning models in open-world and safety-critical scenarios, where test inputs may deviate from the training distribution and overconfident predictions on unknown samples can lead to unreliable decisions. Outlier Exposure (OE) has emerged as a promising OOD detection paradigm by introducing auxiliary outliers during training to enlarge the margin between in-distribution (ID) and OOD samples. Existing OE-based methods typically enlarge this margin by employing uniform labels to maximize the entropy of OOD samples over ID categories. However, we theoretically show that uniform labels inevitably disregard the relations between OOD samples and ID categories, termed the over-softening effect, leading to a suboptimal margin bound. Our theoretical analysis further reveals that explicitly exploiting such relations can instead yield improved OOD detection performance. Motivated by this insight, we propose \underline{A}daptive Confidence \underline{OE} (AOE), a simple yet effective method that leverages temperature scaling to recalibrate outlier labels. Specifically, AOE generates adaptive soft targets from temperature-scaled model predictions for OOD samples, where the learnable temperature smooths the prediction distribution without fully erasing class-wise relational information. By supervising OOD samples with these adaptive soft targets, AOE preserves the semantic proximity between OOD samples and ID categories while encouraging the softened targets to approach a high-entropy distribution, thereby suppressing overconfident OOD predictions and enlarging the separation margin. Extensive experiments across diverse benchmarks demonstrate the effectiveness of AOE.

With the rapid progress of large language models (LLMs), reliably evaluating the capabilities of pre-trained LLMs has become increasingly important. The challenge is that base pre-trained models are optimized for next-token prediction and often fail to follow instructions or produce well-formed answers under standard prompting and direct decoding. As a result, benchmark performance can conflate model capability with decoding-induced failures to produce task-oriented outputs, while exposing such behavior often relies on costly post-training. Recent decodingonly approaches attempt to reshape output distributions, but such methods can be inefficient and brittle across open-ended tasks. To address these limitations, we propose Energy-Based Decoding (EBD), a training-free, reward-guided framework for activating task-oriented behaviors from frozen pre-trained LLMs across both open-ended and objective tasks. EBD augments decoding with an external lightweight reward model, steering generations toward high-utility responses while anchoring them to the pre-trained model prior through a reward-tilted target distribution. We show that EBD shifts base-model outputs toward more instructionfollowing behavior, increasing behavioral similarity to post-trained counterparts and enabling a fairer inference-time evaluation of accessible pre-trained-model behavior. Empirically, EBD outperforms baselines across five models and six benchmarks, improving Qwen3-8B-Base on AlpacaEval2.0 from 8.8 to 44.5, reducing Mistral-7B Math500 latency by 18.9x relative to prior decoding work, and remaining robust to reward-model size.

Given the real-time demands of UAV tracking, many methods simplify the backbone to reduce computation, but this often weakens feature representation and degrades performance in complex scenarios. To alleviate this issue, we propose EATrack, an efficient and asymmetric UAV tracking framework centered around a teacher-guided dual-branch distillation strategy that enhances the feature expressiveness of the lightweight student model. Specifically, EATrack investigates two complementary perspectives of knowledge transfer: spatially focused feature-level distillation that compensates for weakened representations by guiding the student to learn strong target representations, and prediction-level distillation that enhances spatial localization by learning the teacher's capability for accurate target localization. Furthermore, to enhance robustness against appearance variations, we introduce a fine-grained target-aware distillation strategy that selectively transfers the teacher's target modeling capacity to the student. A temporal adaptation module is incorporated at inference to enhance robustness over time. Experiments on five UAV benchmarks demonstrate that EATrack achieves a favorable balance between accuracy and speed. Code: https://github.com/GXNU-ZhongLab/EATrack

Ultra-low-field (ULF) MRI offers portable and low-cost imaging but suffers from poor image quality. To address this, we present our submission to the 2025 ULF Enhancement Challenge (ULF-EnC), where the goal is to synthesise high-field-like MRIs from 64 mT scans. Our pipeline enhances ULF MRI through a combination of anatomical conditioning and model ensembling. We first generate tissue segmentation priors using a Swin UNETR trained solely on challenge-provided data. These priors condition two independent enhancement networks - a CycleGAN and a transformer-based residual enhancement model (T-REX) - each trained to synthesise 3 T-like MRIs. Outputs from both models are combined using a weighted average. Our approach produces enhanced MRIs that were comparable to high-field scans both quantitatively and qualitatively.

On-policy self-distillation (OPSD) improves the reasoning performance of large language models (LLMs) by providing dense token-level supervision for on-policy rollouts. However, existing OPSD methods often yield limited gains on in-domain reasoning and generalize poorly to out-of-domain problems. We identify two key causes: conditioning the self-teacher on a verified solution encourages imitation of training-domain reference trajectories rather than error-specific correction, and applying distillation to the full response can overwrite valid reasoning prefixes and reinforce overfitting. We propose Reflective On-policy Self-Distillation (ROSD), a framework that turns reference-solution imitation into targeted reasoning correction through reflection-guided, error-localized distillation. For each rollout, ROSD uses a self-reflector to extract a corrective idea and locate the first erroneous span. The corrective idea guides the self-teacher toward targeted supervision, while the localized error span restricts distillation to where correction is needed. This design corrects flawed reasoning while preserving valid prefixes. Experiments on multiple in-domain and out-of-domain reasoning benchmarks show that ROSD yields stronger in-domain reasoning performance overall and substantially better out-of-domain generalization than standard OPSD. Code is available at https://github.com/ZiqiZhao1/ROSD.

Multimodal Large Language Models (MLLMs) exacerbate safety risks by introducing vulnerabilities across multiple modalities, such as language and vision. Current MLLM safety evaluation tools, however, suffer from major limitations: 1) English-centric dataset construction, and 2) a focus on generic risks that are not tied to local cultural contexts. This paper introduces KSAFE-MM, a benchmark for Korean multimodal safety evaluation that covers both general safety risks and culture-specific vulnerabilities. KSAFE-MM consists of two parts, KSAFE-MM-G and KSAFE-MM-C. KSAFE-MM-G evaluates globally shared risks in Korean contexts through linguistic contextualization, which transforms generic safety queries into contextually grounded multimodal samples. KSAFE-MM-C targets culture-dependent MLLM safety vulnerabilities using localized visual queries derived from real-world contexts. It pairs these visual queries with jailbreak-style textual queries to cover multimodal safety risks involving cultural visual cues and malicious textual intent. Together, these components provide a general-to-local construction pipeline for evaluating both globally shared safety risks and culture-specific vulnerabilities. We evaluate 12 state-of-the-art MLLMs on KSAFE-MM and reveal that models exhibit greater vulnerability to culturally grounded attacks than to generic ones. Notably, jailbreaking strategies substantially amplify attack success rates, with ProgramExecution yielding up to 74.2% ASR compared to 13.4% for standard queries. Furthermore, we identify a systematic trade-off between safety and over-refusal, where models achieving low ASR tend to exhibit excessive refusal behavior on benign queries. These findings highlight the urgent need for culturally grounded safety evaluation beyond English-centric benchmarks.

Quantifying impact phenomena in badminton smashes is important for evaluating both athletic performance and equipment; however, conventional measurement systems involve trade-offs between temporal resolution, data efficiency, and preparation effort. This study proposes a measurement method using two synchronized event cameras to automatically estimate impact time, impact location on the racket face, and post-impact shuttlecock speed in an integrated manner within the same trial. The swing interval was detected from event rate statistics, impact time was estimated from the shuttlecock trajectory inflection in the lateral-view event data, impact location was determined by ellipse fitting to the racket face in the rear-view event image, and shuttlecock speed was calculated in the sagittal plane. To validate the proposed method, Bland-Altman analysis was performed against a high-speed camera-based reference method using 125 smash trials from five players. Impact time and shuttlecock speed were estimated in all 124 analyzable trials, and impact location was estimated in 93.5% (116/124). The bias (95% CI) for impact time, medio-lateral impact location, longitudinal impact location, and shuttlecock speed were 1.84 ms (1.45 to 2.23), 3.45 mm (2.18 to 4.72), -1.92 mm (-2.97 to -0.88), and -1.00 m/s (-2.46 to 0.46), respectively. No proportional bias was observed for any metric. These results suggest that the proposed method can serve as a useful tool for integrated assessment of badminton smash performance and equipment in practical settings.

Self-evolving large language models (LLMs) learn by generating their own training tasks and solutions, reducing reliance on human-curated supervision. However, in many reasoning domains, the model must also validate generated tasks and judge generated answers to obtain training signals. This creates a training-signal challenge: erroneous self-judgments become erroneous gradient updates. Existing approaches either rely on external verifiers, which limits generality, or treat noisy self-generated feedback as supervision. We propose COSE (Confidence-Orchestrated Self-Evolution), which uses the LLM's intrinsic confidence as a lightweight uncertainty signal to modulate learning. COSE introduces confidence-weighted PPO updates and confidence-prioritized replay. Across 19 held-out benchmarks and four Qwen/Llama backbones (0.6B--4B), COSE consistently improves over base models and achieves the best average performance in general reasoning and mathematics, while remaining competitive on code. Code and data are available at https://anonymous.4open.science/r/COSE_-B5C2.

Memory-augmented large language models extend reasoning beyond a fixed context window by maintaining long-term memory across interactions. However, existing memory systems often collapse stable user facts, episodic events, and behavioral rules into a shared space, allowing functionally distinct memories to be retrieved and used as interchangeable evidence. We identify this failure mode as heterogeneous memory contamination, where context-specific events become overgeneralized claims, or semantically relevant but functionally incompatible memories mislead generation. To this end, we introduce MemGuard, a type-aware memory framework that preserves functional memory boundaries during memory construction and retrieval. It assigns each memory an explicit functional role at write time, maintains relations across type-isolated memories, and selectively composes evidence only from necessary memory types, reducing contamination from irrelevant or functionally incompatible evidence. Across hallucination and long-horizon conversation benchmarks, MemGuard improves memory reliability by up to 28.27% while retrieving up to 5.8x fewer memory tokens than prior methods. These results suggest that reliable long-term reasoning depends on principled organization and selective use of heterogeneous memory.

Large language models (LLMs) can now solve complex problems through long chain-of-thought (CoT) reasoning, but the trade-off between performance and token cost remains a central challenge. To address this issue, supervised fine-tuning (SFT) often uses compressed reasoning data, where CoT traces are shortened into compact forms. However, the effect of such compressed reasoning data on post-training remains poorly understood. In this paper, we propose a taxonomy of CoT consisting of Explicit CoT, which outputs all operations without aggregation, Composed CoT, which combines multiple operations into a single step, and Implicit CoT, which omits intermediate operations. We construct a synthetic compositional reasoning task that allows controlled variation of difficulty, compression granularity, and data size, and conducted a comprehensive set of experiments across different model families and sizes. Notably, we find that (i) coarser CoT requires more SFT data, (ii) compared with Explicit CoT, Composed CoT and Implicit CoT benefit more from data scaling, while Composed CoT benefits from data repetition and Implicit CoT tends to lead to memorization, (iii) unlike SFT, subsequent reinforcement learning (RL) with verifiable rewards (RLVR) decomposes compressed steps learned during SFT, and (iv) unidirectional CoT ordering shows stronger generalization on longer sequential tasks. Our findings provide implications for CoT design under data resource constraints and offer important insights into the mechanisms of SFT and RL in LLM post-training.

Deep networks are powerful function approximators, but they typically store many different computations in shared weight matrices, making it difficult to selectively reuse or adapt parts of them when a familiar structure appears in novel combinations. We introduce the Vector Network (VN), a hierarchical recurrent architecture in which each layer replaces a fixed weight matrix with a library of reusable rank-1 weight atoms. For each input, VN minimizes a layer-local energy to infer a sparse set of active weight atoms and their coefficients, jointly constrained by bottom-up input reconstruction and top-down feedback consistency. These weight atom coefficients then compose an input-specific low-rank weight matrix for that sample. After convergence, slow learning updates only the selected weight atoms through local residual signals scaled by the inferred coefficients. We evaluate VN on four compositional benchmarks spanning 1D signals, 2D spatial decoding, N-body dynamics, and compositional MNIST. VN matches strong baselines in distribution while often achieving out-of-distribution error about an order of magnitude lower when familiar factors must be recombined in novel ways. Vector networks thus make compositional generalization a structural property of the architecture and inference process rather than a brittle byproduct of fitting many behaviors into one shared dense parameter substrate.

Understanding how LLMs reason is hindered by a practical asymmetry: while their generated outputs are observable, the underlying reasoning patterns remain opaque. Relying on single probes, such as Mutual Information Peak (MIP) or Deep-Thinking Ratio (DTR), risks underestimating the genuine inferential structure. To response this deficiency, we present an Integrated, cross-Architecture Reasoning (IAR) framework, designed to provide a unified approach to LLM reasoning interpretability. Specifically, we first propose to use bandwidth-calibrated MIP coupled with Tukey IQR peak-detection to isolate reasoning-crucial tokens at the output layer. Second, we performed an overlap analysis between MIP-picked tokens and DTR-deep tokens to trace the cross-layer trajectories of those tokens. This also discloses whether reasoning-crucial tokens are computation-intensive as well, further facilitating to understand how reasoning patterns evolve across model layers. Finally, we apply a Jaccard stability metric over multi-domain problems to verify if the MIP-identified tokens are reasoning quality-guaranteed. Extensive experiments on three models (Qwen-7B, Qwen-14B, and Llama-8B) across four domains (mathematics, code, logic, and common sense) demonstrate IAR's generalizable interpretation capabilities across architectures.

GraphRAG extends retrieval-augmented generation by organizing corpora as explicit knowledge graphs, enabling graph-based retrieval for complex question answering. However, existing frameworks extract entities and relations within individual chunks, leaving cross-chunk relations -- those whose evidence spans multiple passages -- systematically absent from the index. Exhaustive LLM-based recovery of such relations is impractical due to the combinatorial explosion of chunk combinations. We present CrossAug, a GNN-guided CROSS-Chunk Graph AUGmentation method that enriches GraphRAG indices with cross-chunk relational structure as an offline step before query-time retrieval. CrossAug derives training supervision through self-supervised graph corruption, uses a topology-aware GNN to score subgraphs for missingness, and applies evidence-grounded LLM completion only to selected high-scoring regions. Experiments on three LLM-based GraphRAG frameworks across four multi-hop and long-document QA benchmarks demonstrate that CrossAug consistently improves performance, confirming the benefit of cross-chunk graph augmentation for retrieval-based question answering. Our code is available at https://github.com/DonFinliani/CrossAug.

The frontier of mathematics is defined by problems whose solutions are not yet known, yet it remains unclear whether language models can meaningfully engage with such problems without human intervention. A major obstacle is the lack of large-scale research-level math datasets. To this end, we introduce ResearchMath-14k, a set of $14{,}056$ problems curated from academic sources via a multi-agent pipeline, making it the largest collection of research-level mathematical problems to date. We further generate ResearchMath-Reasoning, $220$K teacher trajectories from two open models, where we observe recurring avoidance behaviors such as non-attempts and fabricated references. Interestingly, across eight open-weight models, newer generations produce $5.6\times$ more references and $5.0\times$ more fake references per trace. After agentic filtering of ResearchMath-Reasoning, fine-tuning Qwen3 models from 4B to 30B parameters improves over base models by $9.2$ points on average. This shows that filtered open-problem attempts can provide useful supervision even without fully correct reasoning traces. We make ResearchMath-14k publicly available for future works on research-level mathematical reasoning.

We empirically audit the k-NAF budget-accounting mechanism in Anchored Decoding using (i) a fixed, class-stratified workload (approximately 8,500 randomized executions across six prompt classes) and (ii) an adaptive prompt-search procedure targeting high proxy spend ratios. On the fixed workload, mean cumulative KL spend remains far below the sequence-level budgets K in {600, 1000}, and an empirical Bernstein-style proxy stays below K for every class; surface-overlap diagnostics (ROUGE-L and 5-gram Jaccard) are correspondingly small. Adaptive search increases the proxy spend ratio but does not produce clear budget exhaustion. On a held-out copyright-domain workload at k = 3, several prompts exhibit proxy ratios above 1 under early-stopped evaluations with small realized sample sizes; re-evaluating the same prompts with larger allocation reduces the proxy ratio to the range [0.26, 0.40] under comparable mean spend, consistent with proxy artifacts rather than per-trajectory budget failures.