spkc83/retail-bank-servicing-agent-9b

Retail Bank Agent 9B

This is a research checkpoint for a synthetic retail-bank customer-service demonstration. It is a merged bf16-lora LoRA adaptation of spkc83/retail-bank-servicing-agent-9b at revision 1d56824995aa1adecfe20f62ca42fb1c0c443817. The model has approximately 8.8 billion parameters and uses the base model's native tagged JSON tool-call format.

Training

• Dataset: spkc83/retail-bank-servicing-alignment-sft at f7784b34b41094b1e771323b2df046ed4664b9a4
• Training records: 1222
• Validation records: 277
• Tool manifest: nine synthetic retail-banking tools
• Assistant-only target masking: tool-call and final-assistant spans
• Maximum sequence length: 2048
• Optimizer steps: 750
• LoRA rank/alpha: 32/64
• Source commit: 312257ee6fb33f4c8ecad38eaea60e59709e607d
• Chat-template SHA-256: 6727ca16a39df05c41af54eb651aa618b50a29967ad3951a31b90c4e385573fc

The released root checkpoint is merged FP16 weights. The trained BF16 adapter is also stored under adapter/.

Intended use and limitations

The model is intended only for the linked synthetic banking POC. It must be given the published tool schemas and tool results. It has no access to real banking systems, is not financial advice, and may make incorrect tool choices or unsupported claims. Evaluate tool-call syntax, arguments, backend execution, grounded final responses, OOD behavior, and multi-turn behavior before relying on a revision.