FeedSources
BR

Baseten Research

7 articles total

Go to source

Serve and scale open-source and custom AI models on the fastest, most reliable inference platform.

Go to source
  •  Can a Language Model Learn Facts Continually in Its Weights?
  •  Post-Training Science for Supervised Fine-Tuning
  •  Still: Amortized KV Cache Compaction in a Single Forward Pass
  •  Post-training frontier legal agents with Baseten Research
  •  Towards infinite context windows: neural KV cache compaction
  •  Dense, on-policy, or both?
  •  Repeated KV cache for long-running agents