Verifiable reasoning training compresses an entire response to a binary pass-or-fail label, yet a fail label alone does not reveal whether the corresponding trajectory should be discarded, penalized, or repaired. Many failed trajectories remain close to the correct solution through substantial portions of their intermediate reasoning and deviate only at a later branching point, whereas others contain no reliable prefix that can support a correct suffix. Treating both failure types as uniformly negative supervision suppresses useful reasoning states while reinforcing erroneous ones. To address this limitation, we propose prima, which decomposes failed trajectories into inheritable prefixes and divergent suffixes, and rewrites repairable failures into shared-prefix preference pairs for localized reasoning optimization. We formalize the optimization cost of mixing the two failure types and show that treating them as a uniform negative class shrinks effective gradient magnitude, providing a principled justification for separating repairable from irrecoverable failures before training. Experiments show that prima converts failed trajectories from coarse-grained negative supervision into cleaner suffix-level supervision signals, leading to consistent gains on code and mathematical reasoning tasks.

Recently, the role of LLM-as-judge in evaluating large language models has gained prominence, emerging as an important method to partially replace costly human assessment. However, current judge models suffer from narrow specialization and limited robustness, undermining their capacity for comprehensive evaluations. In this work, we decompose the ability of a generalist and generative judge model into three levels: objective verification, subjective evaluation, and rubric refinement. We present CompassJudger-2, a novel generalist judge model that overcomes these limitations via a task-driven, multi-scenarios data curation strategy. We conducted large-scale data collection for each type of task and designed tailored rejection sampling strategies to filter the data, ensuring data diversity, accuracy, and effectiveness. Empirically, CompassJudger-2 achieves superior results across multiple judge and reward benchmarks, demostrating its excellent robustness and generalization ability. These contributions advance robust, scalable LLM judgment and establish new performance and evaluation standards.

Post-training large language models usually apply a single training recipe to all samples, even though the model’s own rollouts reveal different sample-level learning states. We propose Self-Routing, a behavior-conditioned post-training framework that uses rollout correctness and confidence to decide how each sample should be optimized. Depending on its behavior state, a sample is routed to GRPO, on-policy self-distillation, regularization, or skipping, allowing training to adapt without external teachers, extra annotations, or additional sampling. Experiments on mathematical reasoning across Qwen3 and Qwen3.5 backbones show that Self-Routing consistently improves over uniform GRPO, uniform OPSD, fixed mixtures, and simpler routing baselines. Further analyses show that the routing distribution changes over training and reduces unnecessary updates on low-signal or already stable samples.

Machine unlearning aims to make a model forget specific data, yet unlearned LLMs often fail to stay unlearned: brief fine-tuning can revive removed knowledge. Existing robustness predictors rely on global weight-space displacement, but distance alone can be misleading when random or destructive updates collapse performance. We argue that relearning robustness depends on update structure: robust unlearning should affect forget-critical weights while sparing retain-critical ones. We introduce the Forget-Retain Alignment Gap (FRAG), a training-free predictor that scores an update’s forget-retain alignment without running a relearning attack, and separates selective from dense updates more reliably than global distance. Building on the forget-critical, retain-sparing principle, Forget-Retain Pruning (FRP) improves relearning robustness. Our results suggest that weight selectivity better explains robustness than distance alone. Our code is available at https://github.com/Yi1-Chen/FRAG.

Video misinformation detection is often approached as a binary veracity classification problem, overlooking the complex reasoning required to explain how and why content misleads. Existing benchmarks fail to capture the diversity of manipulation strategies, such as AI-generated edits and out-of-context manipulation, and do not evaluate whether models can provide process-level justifications for their judgments. We address these limitations with MisVideoQA, a multi-turn benchmark designed to assess comprehensive understanding and reasoning in video misinformation analysis. MisVideoQA covers 12 fine-grained deception categories and evaluates models along six dimensions, progressing from perceptual attribution to intent and persuasion analysis. Recognizing that standard MLLMs struggle to sustain such structured, evidence-based deduction, we propose MisAgent, a Delphi-inspired multi-agent framework in which specialized agents collaboratively integrate multimodal cues with external evidence. Experimental results show that state-of-the-art multimodal large language models perform poorly on MisVideoQA, while MisAgent consistently improves reasoning accuracy and explanation quality. Together, our benchmark and framework establish a unified foundation for reliable, interpretable, and evidence-grounded video misinformation analysis.

Social media platforms provide an ideal environment to spread misinformation, where social bots can accelerate the spread. This paper explores the interplay between social bots and misinformation on the Sina Weibo platform. We construct a large-scale dataset that includes annotations for both misinformation and social bots. From the misinformation perspective, the dataset is multimodal, containing 11,393 pieces of misinformation and 16,416 pieces of verified information. From the social bot perspective, this dataset contains 65,749 social bots and 345,886 genuine accounts, annotated using a weakly supervised annotator. Extensive experiments demonstrate the comprehensiveness of the dataset, the clear distinction between misinformation and real information, and the high quality of social bot annotations. Further analysis illustrates that: (i) social bots are deeply involved in information spread; (ii) misinformation with the same topics has similar content, providing the basis of echo chambers, and social bots would amplify this phenomenon; and (iii) social bots generate similar content aiming to manipulate public opinions.

While recent advances in fake news video detection have shown promising potential, existing approaches typically (1) focus on a specific domain (e.g., politics) and (2) assume the availability of multiple modalities, including video, audio, description texts, and related images. However, these methods struggle to generalize to real-world scenarios, where questionable information spans diverse domains and is often modality-incomplete due to factors such as upload degradation or missing metadata. To address these challenges, we introduce two real-world multi-domain news video benchmarks that reflect modality incompleteness and propose IMOL, an incomplete-modality-tolerant learning framework for multi-domain fake news video detection. Inspired by cognitive theories suggesting that humans infer missing modalities through cross-modal guidance and retrieve relevant knowledge from memory for reference, IMOL employs a hierarchical transferable information integration strategy. This consists of two key phases: (1) leveraging cross-modal consistency to reconstruct missing modalities and (2) refining sample-level transferable knowledge through cross-sample associative reasoning. Extensive experiments demonstrate that IMOL significantly enhances the performance and robustness of multi-domain fake news video detection while effectively generalizing to unseen domains under incomplete modality conditions.

The swift detection of multimedia fake news has emerged as a crucial task in combating malicious propaganda and safeguarding the security of the online environment. While existing methods have achieved commendable results in modeling entity-level inconsistency, addressing event-level inconsistency following the inherent subject-predicate logic of news and robustly learning news representations from poor-quality news samples remain two challenges. In this paper, we propose an Event-diven fake news detection framework (Event-Radar) based on multi-view learning, which integrates visual manipulation, textual emotion and multimodal inconsistency at event-level for fake news detection. Specifically, leveraging the capability of graph structures to capture interactions between events and parameters, Event-Radar captures event-level multimodal inconsistency by constructing an event graph that includes multimodal entity subject-predicate logic. Additionally, to mitigate the interference of poor-quality news, Event-Radar introduces a multi-view fusion mechanism, learning comprehensive and robust representations by computing the credibility of each view as a clue, thereby detecting fake news. Extensive experiments demonstrate that Event-Radar achieves outstanding performance on three large-scale fake news detection benchmarks. Our studies also confirm that Event-Radar exhibits strong robustness, providing a paradigm for detecting fake news from noisy news samples.