AI SOTA Benchmarks

Sleep-time Compute: Beyond Inference Scaling at Test-time

Generate, but Verify: Reducing Hallucination in Vision-Language Models with Retrospective Resampling

Perception Encoder: The best visual embeddings are not at the output of the network

It seems like the text you provided is already in English and is a title or a heading for a scientific or technological paper. However, if you have a Chinese text that you would like translated into English, please provide it, and I will be happy to assist you.

FreshStack: Building Realistic Benchmarks for Evaluating Retrieval on Technical Documents

VistaDPO: Video Hierarchical Spatial-Temporal Direct Preference Optimization for Large Video Models

Exploring Expert Failures Improves LLM Agent Tuning

Set You Straight: Auto-Steering Denoising Trajectories to Sidestep Unwanted Concepts

NoisyRollout: Reinforcing Visual Reasoning with Data Augmentation

It seems like you've provided a term, "Antidistillation Sampling," rather than a full news article or academic report. Could you please provide the complete text you would like translated? This will help me give you an accurate and contextually appropriate translation.

Retrieval-Augmented Generation with Conflicting Evidence

Packing Input Frame Context in Next-Frame Prediction Models for Video Generation

HLS-Eval: A Benchmark and Framework for Evaluating LLMs on High-Level Synthesis Design Tasks

SHeaP: Self-Supervised Head Geometry Predictor Learned via 2D Gaussians

Cobra: Efficient Line Art COlorization with BRoAder References

Towards Learning to Complete Anything in Lidar

Robust and Fine-Grained Detection of AI Generated Texts

BitNet b1.58 2B4T Technical Report

The Scalability of Simplicity: Empirical Analysis of Vision-Language Learning with a Single Transformer

Genius: A Generalizable and Purely Unsupervised Self-Training Framework For Advanced Reasoning

How Instruction and Reasoning Data shape Post-Training: Data Quality through the Lens of Layer-wise Gradients

Seedream 3.0 Technical Report

Heimdall: test-time scaling on the generative verification

Pixel-SAIL: Single Transformer For Pixel-Grounded Understanding

TextArena

xVerify: Efficient Answer Verifier for Reasoning Model Evaluations

Uncertainty-Guided Coarse-to-Fine Tumor Segmentation with Anatomy-Aware Post-Processing

DeepMath-103K: A Large-Scale, Challenging, Decontaminated, and Verifiable Mathematical Dataset for Advancing Reasoning

The Urban Impact of AI: Modeling Feedback Loops in Next-Venue Recommendation

MineWorld: a Real-Time and Open-Source Interactive World Model on Minecraft

PixelFlow: Pixel-Space Generative Models with Flow

Do PhD-level LLMs Truly Grasp Elementary Addition? Probing Rule Learning vs. Memorization in Large Language Models

CoRAG: Collaborative Retrieval-Augmented Generation

SAEs Can Improve Unlearning: Dynamic Sparse Autoencoder Guardrails for Precision Unlearning in LLMs

It seems like the text you provided is already in English. However, if you meant to translate a Chinese title or description related to FlexIP, please provide the Chinese text, and I will be happy to translate it for you. Here’s an example of how the translation might look if the text were in Chinese: **Chinese:** FlexIP: 定制图像生成的动态保存与个性控制 **English:** FlexIP: Dynamic Control of Preservation and Personality for Customized Image Generation If you have a different piece of text that needs translation, please share it!

InteractVLM: 3D Interaction Reasoning from 2D Foundational Models

In-2-4D: Inbetweening from Two Single-View Images to 4D Generation

SQL-R1: Training Natural Language to SQL Reasoning Model By Reinforcement Learning

ModernBERT or DeBERTaV3? Examining Architecture and Data Influence on Transformer Encoder Models Performance

Seaweed-7B: Cost-Effective Training of Video Generation Foundation Model

GigaTok: Scaling Visual Tokenizers to 3 Billion Parameters for Autoregressive Image Generation

Visual Chronicles: Using Multimodal LLMs to Analyze Massive Collections of Images

VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Model

Pangu Ultra: Pushing the Limits of Dense Large Language Models on Ascend NPUs

Training-free Guidance in Text-to-Video Generation via Multimodal Planning and Structured Noise Initialization

ZipIR: Latent Pyramid Diffusion Transformer for High-Resolution Image Restoration

Latent Diffusion Autoencoders: Toward Efficient and Meaningful Unsupervised Representation Learning in Medical Imaging

SocioVerse: A World Model for Social Simulation Powered by LLM Agents and A Pool of 10 Million Real-World Users

InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models

LLM Can be a Dangerous Persuader: Empirical Study of Persuasion Safety in Large Language Models

S1-Bench: A Simple Benchmark for Evaluating System 1 Thinking Capability of Large Reasoning Models

LLM-SRBench: A New Benchmark for Scientific Equation Discovery with Large Language Models

It seems like the text you provided is already in English. However, if you meant to translate a Chinese title or description of "Mavors: Multi-granularity Video Representation for Multimodal Large Language Model," please provide the Chinese text, and I will be happy to translate it for you.

It seems like the text you provided is already in English. However, if you meant to translate a title or a concept from Chinese to English, please provide the Chinese text, and I will be happy to translate it for you. If you have any other news or academic achievement reports in Chinese that need translation, feel free to share them!

M1: Towards Scalable Test-Time Compute with Mamba Reasoning Models

Breaking the Data Barrier -- Building GUI Agents Through Task Generalization

Towards Automated Safety Requirements Derivation Using Agent-based RAG

C-SHAP for time series: An approach to high-level temporal explanations