ReVonance Support Drafter (SFT, lora)
The Drafter agent of a LangGraph multi-agent support system for the fictional shop ReVonance.
Base model TinyLlama/TinyLlama-1.1B-Chat-v1.0, fine-tuned with LoRA (merged).
Stage: sft (sft = supervised on the response simulator; dpo = then aligned with human and
critic preferences via Direct Preference Optimisation).
Evaluation (held-out simulated tickets, full multi-agent graph)
| stage | method | temperature | first-draft approval | fda 95% CI | resolved | avg drafts |
|---|---|---|---|---|---|---|
| base | none | 0 | 0.006 | 0.00–0.03 | 0.006 | 2.988 |
| sft | lora | 0 | 0.988 | 0.96–1.00 | 0.988 | 1.024 |
| sft | lora | 0.7 | 1 | 0.98–1.00 | 1 | 1 |
Knowledge-base fingerprint: f9cdb07b5df6. Demo model trained on synthetic data; it only
knows the toy policies of ReVonance.