2 papers
cs.CL2026
DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression
DeepSeek-AI, :, Anyi Xu +585
The widespread adoption of long-horizon agents has made model workloads increasingly input-heavy. Although prior work has substantially reduced the cost of long-context computation…
cs.AI2026
ThinkFlow: Self-Evolving Probabilistic Latent Memory for Lifelong Conversational Agents
Cai Ke, Xin Liu, Han Zhang +6
Lifelong conversational agents rely on memory systems to maintain deep, context-aware interactions with users. However, existing explicit textual memory pipelines suffer from a sev…