ByteDance Seed team's VINCIE-NExT transfers image editing to video by editing keyframes and using them as visual references, achieving 3.08 on OpenVE-Bench with TDF3D-RoPE positional encoding and Chain-of-Editing inference, though camera viewpoint edits remain challenging.
ByteDance's Seed team discovered that DeepSeek-V4 models exhibit periodic accuracy fluctuations every 4 tokens in long-context retrieval, caused by chunked KV cache compression creating phase sensitivity where retrieval performance depends on information position relative to compression window boundaries, a phenomenon mitigated but not eliminated by post-training.
ByteDance Seed and TokenWave introduce three benchmarks—ASPIRE, S³Gym, and HarnessDev—to evaluate whether AI agents can autonomously form goals, learn from experience, and retain improvements, showing that current agents struggle to translate self-assessment into lasting capability gains.
ByteDance Seed and TokenWave introduce ASPIRE, S³Gym, and HarnessDev — three benchmarks that test whether AI agents can autonomously select learning goals, distill experience into improved decisions, and persistently upgrade their own execution systems without human-provided verification.
ByteDance Seed and collaborators introduce ASPIRE, S³Gym, and HarnessDev benchmarks to study how agents learn from vague goals, self-evaluate actions, and persist improvements, revealing that current agents overfit to proxy feedback and fail to convert self-judgment into lasting capability gains.
