PEFT
Safetensors
English
decision-model
system-1
system-one
calibration
tool-use
agents
guardrails
prompt-injection
lora

Commit History

29 days ago

  1. Version 2: agent steps, prompt injection, grounding, multi-turn call/ask/decline

    Verified signature
    281260f

  1. Describe uses beyond the cascade: hooks, guardrails, trace labelling

    Verified signature
    47bd094
  2. Link the 3B sibling and the collection

    Verified signature
    d670499
  3. adapter_config: task_type CAUSAL_LM

    Verified signature
    5aa3ba5
  4. adapter_config: task_type FEATURE_EXTRACTION (Hub requires a string)

    Verified signature
    7fc0a50
  5. Link the GitHub repo and test data

    Verified signature
    c20c42c
  6. agent-decision-9b: LoRA adapter + calibrated joint head for Cloudflare/clef-flash

    Verified signature
    486d484
  7. initial commit

    Verified signature
    71bacea