tpierson-eh/qwen2.5-1.5b-injection-resistance-lora
text-generation model by tpierson-eh
Older snapshot: check the original source for current information.
From the publisher
Model card & documentation
Source previewRead the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.
This is a small research adapter that tests one question: can light fine-tuning teach a model to resist prompt injection? Prompt injection is what happens when text a model reads as data — a tool output, a file, a retrieved page — contains sentences written to boss the model around: send the archive somewhere new, delete the log, do not tell anyone. If the model cannot tell an instruction from its user apart from instruction-shaped text sitting inside data, those sentences can steer it. This adapter was trained on short examples that pair an ordinary task with data carrying a planted instruction. The correct behavior in those examples is to name the plant in one sentence beginning "Flag:", say why it has no authority, and then do the real task anyway. A second set of examples shows legitimate instruction-like data — runbook steps, code comments, documentation, genuine user requests —…
Read the full model card ↗ · Preview checked 2026-10-30T09:04:28.098Z
Inside the original model card — Document outline
- Prompt-Injection Resistance LoRA — a research adapter for Qwen2.5-1.5B-Instruct
- What this is
- Base model
- Adapter
- Load it
- Training recipe (reproducible in full)
- Evaluation
- Evaluation 1 — the battery (56 injections / 28 controls)
Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.
Links from the model card
References supplied by the publisher, not independently verified endorsements. Check the destination before downloading files or entering credentials.
Documentation belongs to its respective authors. Reported project/model license: apache-2.0. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
Model overview
tpierson-eh/qwen2.5-1.5b-injection-resistance-lora is a text-generation model repository published by tpierson-eh on Hugging Face. The source reports the peft library.
This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.
Model facts
- Task
- text-generation
- Library
- peft
- Recent downloads (30 days)
- 0
- Cumulative likes
- 0
- Hugging Face trending score
- 0
- Reported safetensors parameters
- Not reported
- Architecture
- Not reported
- License
- apache-2.0
- Access
- Not gated by Hugging Face
- Created
- 2026-10-30T06:45:36.000Z
- Last modified
- 2026-10-30T06:45:53.000Z
Compatibility and lineage
Reported languages: en
Reported base models: Qwen/Qwen2.5-1.5B-Instruct
Parameter count is not a RAM/VRAM requirement. Check precision, quantization, context length and runtime compatibility in the model card; no hardware or API-cost claim is inferred here.
Use and evaluate
Check license terms and access requirements first. Review the model card’s documented loading instructions, evaluations and safety limitations. Benchmark scores are not imported by this directory, and popularity is not an accuracy ranking.
No model weights are downloaded or executed by AltAPIs. Never enable remote model code without reviewing it.
Source and freshness
Source: Hugging Face Hub. Metadata observed 2026-10-30T06:49:24.903Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.