| --- |
| language: |
| - en |
| license: other |
| license_name: secjev-use-and-distribution-agreement-v1.0 |
| license_link: LICENSE |
| base_model: Qwen/Qwen3.5-4B-Base |
| tags: |
| - security |
| - decision-model |
| - lora |
| - prompt-injection |
| --- |
| |
| # SecJev-4B-Joint-Redo |
|
|
| Redo fine-tune of SecJev-4B for Joint Spot 1 + Spot 2 (single checkpoint, two questions). This repository root is a complete |
| Kev checkpoint directory (LoRA adapter + trained head + tokenizer), loadable |
| with the Kev loader. The Qwen3.5-4B-Base backbone is not re-hosted; |
| `bootstrap.py` fetches the pinned backbone and Kev source. |
|
|
| ## Training |
|
|
| - Compute: DigitalOcean GPU droplet `jev-train-1` (NVIDIA H100 80GB), 2026-11-03 23:03:28–23:37:49 UTC. |
| - Method: `ft_action.py`, 1 epoch, lr 3e-5, batch 32 questions, token budget |
| 4096, gradient checkpointing, fp32 weights with bf16 autocast. |
| Trainable parameters: 33,776,128 (LoRA + head); backbone frozen. |
| - Data: `data/bench/train_joint_t.jsonl` (train) / `data/bench/dev_joint_t.jsonl` (dev), jev-stageA redo bench sets. |
| - Full training log: `redo-training-log.jsonl` and `redo-training.stdout.txt` |
| in this repository. Machine-readable details: `redo-provenance.json`. |
|
|
| ## Dev results (only evaluation so far) |
|
|
| Corrected per-question re-evaluation of this checkpoint on the dev file |
| (score: authorization = P("2") + 0.5·P("1"); injection = P("true")): |
|
|
| | question | dev n | dev pos | AUC | TPR@0.5 | FPR@0.5 | |
| |---|---|---|---|---|---| |
| | authorization | 500 | 85 | 0.9898 | 0.9412 | 0.0048 | |
| | injection | 500 | 250 | 0.9314 | 0.7760 | 0.0200 | |
|
|
| No frozen-test, adaptive-attack, or independent-replication result exists yet |
| for this checkpoint. Dev rows were used for training-time monitoring. Treat |
| the numbers above as development measurements, not deployment guarantees. |
|
|
| Note: the training-time dev metrics printed in `redo-training-log.jsonl` for |
| injection-question runs used an authorization-only scoring formula and are |
| invalid for those runs (they show AUC 0.0 / 0.14); the table above supersedes |
| them. Training itself used correct targets. |
|
|
