1
2
3
Teacher Supervision over Representation Equivalence Classes
颠覆知识蒸馏常规认知：预训练表征只有等价类意义，匹配坐标是伪命题
arXiv:2607.03572v1 Announce Type: cross Abstract: Knowledge distillation is usually framed as a choice of what to match in the teacher - its logits, h…
4
Steerable Visual Representations
最新ECCV 2026录用论文，探索如何对预训练视觉Transformer的表示进行可控操作，为视觉AI带来新的灵活性。
arXiv:2604.02327v2 Announce Type: replace Abstract: Pretrained Vision Transformers (ViTs) such as DINOv2 and MAE provide generic image features that c…
5
Test-Time Training with Next-Token Prediction
挑战测试时训练新范式：用语言模型自监督的下一词预测信号作为内循环目标，无需额外损失函数。
arXiv:2606.21803v1 Announce Type: new Abstract: Next-token prediction is the self-supervised signal that trains language models, and every observed pr…
6
Mordal: Automated Pretrained Model Selection for Vision Language Models
自动化选择视觉语言模型预训练方案，提升多模态任务效率
arXiv:2502.00241v2 Announce Type: replace-cross Abstract: Incorporating multiple modalities into large language models (LLMs) is a powerful way to enh…
7
Let ViT Speak: Generative Language-Image Pre-training
ViT也能开口说话？全新生成式语言-图像预训练框架，让视觉与语言深度融合！
arXiv:2605.00809v2 Announce Type: replace Abstract: In this paper, we present \textbf{Gen}erative \textbf{L}anguage-\textbf{I}mage \textbf{P}re-traini…
8
Leveraging Metric Depth for Relative Depth Prediction
利用预训练模型零样本能力，解决足球场景下训练样本少的单目深度估计难题，方法新颖且实用。
arXiv:2606.10628v1 Announce Type: new Abstract: We present our solution to the 2025 SoccerNet Monocular Depth Estimation Competition Challenge. Predic…
9
SindBERT, the Sailor: Charting the Seas of Turkish NLP
专为土耳其语优化的BERT模型SindBERT，填补低资源语言NLP空白
arXiv:2510.21364v2 Announce Type: replace Abstract: Transformer models have revolutionized NLP, yet many morphologically rich languages remain underre…
10
Elastic ViTs from Pretrained Models without Retraining
无需额外训练，即可让预训练ViT弹性适配不同计算约束，实现高效推理与精度平衡的创新方法。
arXiv:2510.17700v2 Announce Type: replace Abstract: Vision foundation models achieve remarkable performance but are only available in a limited set of…
11
Goldfish: Monolingual Language Models for 350 Languages
覆盖350种语言的单语模型Goldfish，大幅提升低资源语言NLP能力。
arXiv:2408.10441v3 Announce Type: replace Abstract: For many low-resource languages, the only available language models are large multilingual models …
12
The Missing Piece in Pre-trained Model Evaluation: Reward-Guided Decoding Unlocks Task-Oriented Behavior Without Parameter Updates
提出奖励引导解码框架，让预训练模型无需参数更新即可激发面向任务行为，补齐评估环节的缺失拼图。
arXiv:2605.28020v1 Announce Type: new Abstract: With the rapid progress of large language models (LLMs), reliably evaluating the capabilities of pre-t…
13
Architecture-driven Shift: towards a lightweight selector for capturing the trends of logit shift
提出轻量级选择器捕捉logit shift趋势，高效平衡持续学习中的“可塑-稳定”矛盾
arXiv:2605.27469v1 Announce Type: cross Abstract: Continual Learning (CL) is a practical paradigm to utilize power of deep pre-trained neural networks…
14
SMART Fine-tuning Factor Augmented Neural Lasso
提出SMART框架，将预训练模型融入高维非参数变量选择，为微调提供理论基础。
arXiv:2604.12288v2 Announce Type: replace-cross Abstract: Fine-tuning is a widely used strategy for adapting pre-trained models to new tasks, yet its …
15
From BERT to T5: A Study of Named Entity Recognition
从BERT到T5，一篇扎实的NER微调实战对比，技术细节丰富。
arXiv:2605.18462v1 Announce Type: new Abstract: Named entity recognition (NER) has been one of the essential preliminary steps in modern NLP applicati…
16
Rethinking 1-bit Optimization Leveraging Pre-trained Large Language Models
颠覆传统，用预训练大模型突破1-bit量化瓶颈，既省存储又保精度。
arXiv:2508.06974v2 Announce Type: replace Abstract: 1-bit LLM quantization offers significant advantages in reducing storage and computational costs. …