DeepSeek-V-4-Flash

The 'DeepSeek-V-4-Flash' dataset/benchmark is used to evaluate the performance of language models in the context of autonomous training and post-training processes.

Papers using DeepSeek-V-4-Flash (3)

AutoTrainess: Teaching Language Models to Improve Language Models Autonomously2026SETA: Scaling Environments for Terminal Agents2026Bayesian-Agent: Posterior-Guided Skill Evolution for LLM Agent Harnesses2026