Blame for README.md

---
base_model: Qwen/Qwen2.5-7B-Instruct
library_name: peft
license: other
---
# incidentbench-binary_cond

LoRA adapter (r=16) on Qwen2.5-7B-Instruct for the IncidentBench joint classifier: one answer from seven MIT risk domains or `No incident (potential only)`. Detection and category are both read from that one answer.

Question this arm answers: Exp 4 control: binary probe with the conditional 'If this becomes an incident' wording.

Training labels are model-written ("silver") under the project's annotation guideline, not human gold. Research artefact: scores and known failure modes (it under-calls incidents) are in the IncidentBench report; do not use it for decisions about real systems.

Prompt format, training rows, and hashes are in `data_manifest.json`; the training recipe is in `train_manifest.json`. The adapter is applied unmerged on top of the base model (as in every reported run). Serve with vLLM: `vllm serve Qwen/Qwen2.5-7B-Instruct --enable-lora --max-lora-rank 16 --lora-modules binary_cond=chiffonng/incidentbench-binary_cond` and send the arm's prompt (train/build_data.py transform `binary_cond`), greedy decoding. Training stack: transformers 5.9.0, peft 0.19.1.