README.md

---
base_model: google/gemma-3-27b-it
library_name: peft
license: gemma
pipeline_tag: text-generation
tags:
- lora
- peft
- gemma3
- information-extraction
---

# General-IE Baseline Adapter — NER (Gemma 3 27B)

LoRA adapter for `google/gemma-3-27b-it` trained with an **IEInstruct-style named-entity-recognition only** objective. Released as a *general information-extraction baseline* for the anonymous submission *"From Lengthy Narrative to Structured Data: Instruction Fine-Tuning Open-Weight LLMs for Information Extraction from Corporate Disclosures."*

On the paper's compensation-consultant task this generic-IE adapter reaches **F1 75.8%**, well below the domain-specific adapters (up to 96.1%) — evidence that task-specific fine-tuning outperforms generic IE instruction-tuning for this domain.

| | |
|---|---|
| Base model | `google/gemma-3-27b-it` |
| Method | LoRA (r=8, α=16), 4-bit QLoRA |
| Training objective | named-entity-recognition only |
| Instance-level F1 (consultant task) | **75.8%** |

## Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

base = "google/gemma-3-27b-it"
tok = AutoTokenizer.from_pretrained(base)
model = AutoModelForCausalLM.from_pretrained(base, device_map="auto", load_in_4bit=True)
model = PeftModel.from_pretrained(model, "cs-file-uploads/ie-ner-adapter")
```

## License

Derived from Google **Gemma 3**; use is subject to the [Gemma Terms of Use](https://ai.google.dev/gemma/terms). Adapter weights are released for research use.

## Citation

```bibtex
@misc{anonymous2026fromlengthy,
  title={From Lengthy Narrative to Structured Data: Instruction Fine-Tuning Open-Weight LLMs for Information Extraction from Corporate Disclosures},
  author={Anonymous},
  year={2026},
  note={Under review}
}
```