The promising zero-shot generalization of vision-language models such as CLIP has led to their adoption using prompt learning for numerous downstream tasks. Previous works have shown test-time prompt tuning using entropy minimization to adapt text prompts for unseen domains. While effective, this overlooks the key cause for performance degradation to unseen domains - distribution shift. In this work, we explicitly handle this problem by aligning the out-of-distribution (OOD) test sample statistics to those of the source data using prompt tuning. We use a single test sample to adapt multi-modal prompts at test time by minimizing the feature distribution shift to bridge the gap in the test domain. Evaluating against the domain generalization benchmark, our method improves zero-shot top1 accuracy beyond existing prompt-learning techniques, with a 3.08%improvement over the baseline MaPLe. In cross-dataset generalization with unseen categories across 10 datasets, our method improves consistently across all datasets compared to the existing state-of-the-art.

TPT does not explicitly align the pre-trained CLIP to become aware of the test sample distribution. For the effective test-time adaptation of V -L foundation models, it is crucial to bridge the distribution gap between the pre-training dataset and the downstream evaluation set for high zero-shot generalization.

Large language models (LLMs) are increasingly adopted in clinical decision support, yet aligning them with the multifaceted reasoning pathways of real-world medicine remains a major challenge. Using more than 8,000 infertility treatment records, we systematically evaluate four alignment strategies: Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), Group Relative Policy Optimization (GRPO), and In-Context Learning (ICL) through a dual-layer framework combining automatic benchmarks with blinded doctor-in-the-loop assessments. GRPO achieves the highest algorithmic accuracy across multiple decision layers, confirming the value of reinforcement-based optimization for structured prediction tasks. However, clinicians consistently prefer the SFT model, citing clearer reasoning processes (p = 0.035) and higher therapeutic feasibility (p = 0.019). In blinded pairwise comparisons, SFT attains the highest winning rate (51.2%), outperforming both GRPO (26.2%) and even physicians' original decisions (22.7%). These results reveal an alignment paradox: algorithmic improvements do not necessarily translate into higher clinical trust, and may diverge from human-centered preferences. Our findings highlight the need for alignment strategies that prioritize clinically interpretable and practically feasible reasoning, rather than solely optimizing decision-level accuracy.

Recent generative models have significantly advanced speech restoration tasks, yet their training objectives often misalign with human perceptual preferences, resulting in suboptimal quality. While post-training alignment has proven effective in other generative domains like text and image generation, its application to generative speech restoration remains largely under-explored. This work investigates the challenges of applying preference-based post-training to this task, focusing on how to define a robust preference signal and curate high-quality data to avoid reward hacking. To address these challenges, we propose a multi-metric preference alignment strategy. We construct a new dataset, GenSR-Pref, comprising 80K preference pairs, where each chosen sample is unanimously favored by a complementary suite of metrics covering perceptual quality, signal fidelity, content consistency, and timbre preservation. This principled approach ensures a holistic preference signal. Applying Direct Preference Optimization (DPO) with our dataset, we observe consistent and significant performance gains across three diverse generative paradigms: autoregressive models (AR), masked generative models (MGM), and flow-matching models (FM) on various restoration benchmarks, in both objective and subjective evaluations. Ablation studies confirm the superiority of our multi-metric strategy over single-metric approaches in mitigating reward hacking. Furthermore, we demonstrate that our aligned models can serve as powerful ''data annotators'', generating high-quality pseudo-labels to serve as a supervision signal for traditional discriminative models in data-scarce scenarios like singing voice restoration. Demo Page:https://gensr-pref.github.io

We introduce SeeingSounds, a lightweight and modular framework for audio-to-image generation that leverages the interplay between audio, language, and vision-without requiring any paired audio-visual data or training on visual generative models. Rather than treating audio as a substitute for text or relying solely on audio-to-text mappings, our method performs dual alignment: audio is projected into a semantic language space via a frozen language encoder, and, contextually grounded into the visual domain using a vision-language model. This approach, inspired by cognitive neuroscience, reflects the natural cross-modal associations observed in human perception. The model operates on frozen diffusion backbones and trains only lightweight adapters, enabling efficient and scalable learning. Moreover, it supports fine-grained and interpretable control through procedural text prompt generation, where audio transformations (e.g., volume or pitch shifts) translate into descriptive prompts (e.g., "a distant thunder") that guide visual outputs. Extensive experiments across standard benchmarks confirm that SeeingSounds outperforms existing methods in both zero-shot and supervised settings, establishing a new state of the art in controllable audio-to-visual generation.

Existing work has shown that o1-level performance can be achieved with limited data distillation, but most existing methods focus on unidirectional supervised fine-tuning (SFT), overlooking the intricate interplay between diverse reasoning patterns. In this paper, we construct r1k, a high-quality reverse reasoning dataset derived by inverting 1,000 forward examples from s1k, and examine how SFT and Direct Preference Optimization (DPO) affect alignment under bidirectional reasoning objectives. SFT on r1k yields a 1.6%--6.8% accuracy improvement over s1k across evaluated benchmarks. However, naively mixing forward and reverse data during SFT weakens the directional distinction. Although DPO can partially recover this distinction, it also suppresses less preferred reasoning paths by shifting the probability mass toward irrelevant outputs. These findings suggest that mixed reasoning data introduce conflicting supervision signals, underscoring the need for robust and direction-aware alignment strategies.

As Large Language Models (LLMs) become increasingly integrated into real-world applications, ensuring their outputs align with human values and safety standards has become critical. The field has developed diverse alignment approaches including traditional fine-tuning methods (RLHF, instruction tuning), post-hoc correction systems, and inference-time interventions, each with distinct advantages and limitations. However, the lack of unified evaluation frameworks makes it difficult to systematically compare these paradigms and guide deployment decisions. This paper introduces a multi-dimensional evaluation of alignment techniques for LLMs, a comprehensive evaluation framework that provides a systematic comparison across all major alignment paradigms. Our framework assesses methods along four key dimensions: alignment detection, alignment quality, computational efficiency, and robustness. Through experiments across diverse base models and alignment strategies, we demonstrate the utility of our framework in identifying strengths and limitations of current state-of-the-art models, providing valuable insights for future research directions.

To understand and identify the unprecedented risks posed by rapidly advancing artificial intelligence (AI) models, this report presents a comprehensive assessment of their frontier risks. Drawing on the E-T-C analysis (deployment environment, threat source, enabling capability) from the Frontier AI Risk Management Framework (v1.0) (SafeWork-F1-Framework), we identify critical risks in seven areas: cyber offense, biological and chemical risks, persuasion and manipulation, uncontrolled autonomous AI R\&D, strategic deception and scheming, self-replication, and collusion. Guided by the "AI-$45^\circ$ Law," we evaluate these risks using "red lines" (intolerable thresholds) and "yellow lines" (early warning indicators) to define risk zones: green (manageable risk for routine deployment and continuous monitoring), yellow (requiring strengthened mitigations and controlled deployment), and red (necessitating suspension of development and/or deployment). Experimental results show that all recent frontier AI models reside in green and yellow zones, without crossing red lines. Specifically, no evaluated models cross the yellow line for cyber offense or uncontrolled AI R\&D risks. For self-replication, and strategic deception and scheming, most models remain in the green zone, except for certain reasoning models in the yellow zone. In persuasion and manipulation, most models are in the yellow zone due to their effective influence on humans. For biological and chemical risks, we are unable to rule out the possibility of most models residing in the yellow zone, although detailed threat modeling and in-depth assessment are required to make further claims. This work reflects our current understanding of AI frontier risks and urges collective action to mitigate these challenges.