Large Reasoning Models (LRMs) have recently demonstrated impressive performance across a range of reasoning tasks by generating intermediate thoughts. However, these models can suffer from overthinking—generating excessive tokens that contribute little to final accuracy while increasing inference cost. To mitigate this, we propose TIV (Thought Injection via Vectors), an innovative framework that compresses token-level reasoning into compact vectors without sacrificing performance. Rather than generating explicit thoughts, TIV injects learnable vectors into the post-attention hidden states of the final token across Transformer layers, enabling implicit and lightweight reasoning. We further introduce a two-stage reinforcement learning strategy: the first stage calibrates the model's reasoning distribution, and the second distills it into a vector-based policy optimized for both accuracy and brevity. Experiments on three reasoning benchmarks show that TIV preserves over 99% of the original accuracy while reducing output length by more than 65% on average, reaching up to 80% in some cases. Moreover, TIV consistently achieves superior trade-offs between accuracy and efficiency compared to existing methods, distinguishing itself as a state-of-the-art (SOTA) approach for efficient reasoning in LRMs.

Facial Emotion Analysis (FEA) extends traditional facial emotion recognition by incorporating explainable, fine-grained reasoning. The task integrates three subtasks—emotion recognition, facial Action Unit (AU) recognition, and AU-based emotion reasoning—to jointly model affective states. While recent approaches leverage Vision-Language Models (VLMs) and achieve promising results, they face two critical limitations: (1) hallucinated reasoning, where VLMs generate plausible but inaccurate explanations due to insufficient emotion-specific knowledge; and (2) misalignment between emotion reasoning and recognition, caused by fragmented connections between observed facial features and final labels. We propose Facial-R1, a three-stage alignment framework that effectively addresses both challenges with minimal supervision. First, we employ instruction fine-tuning to establish basic emotional reasoning capability for reducing hallucinations. Second, we introduce reinforcement training guided by emotion and AU labels as reward signals, which explicitly aligns the generated reasoning process with the predicted emotion. Third, we design a data synthesis pipeline that iteratively leverages the prior stages to expand the training dataset, enabling scalable self-improvement of the model. Built upon this framework, we introduce FEA-20K, a benchmark dataset comprising 17,737 training and 1,688 test samples with fine-grained emotion analysis annotations. Extensive experiments across eight standard benchmarks demonstrate that Facial-R1 achieves state-of-the-art performance in FEA, with strong generalization and robust interpretability.

Large Vision-Language Models (LVLMs) have shown impressive performance across diverse multimodal tasks. However, their trustworthiness is challenged by two types of hallucinations: over-reliance on the language priors of the underlying Large Language Model (LLM) and misalignment between visual and textual modalities. Most existing hallucination mitigation methods often require high training costs, limiting their effectiveness and robustness. To address this, we propose Dual Contrastive Decoding (DCD), which mitigates hallucinations stemming from language priors by subtracting image-irrelevant distributions and further aligns image-text modalities by enabling LVLMs to focus on correct image entities. This strategy requires no additional training and can be applied during the decoding phase of any LVLM, effectively improving the inference capability of LVLMs at minimal cost. Extensive experiments on discriminative (POPE, MME) benchmarks demonstrate that DCD significantly mitigates hallucinations at both the object and attribute levels, while also enhancing overall model performance on general vision-language perception and recognition tasks. The code and dataset will be made available.