README.md

---
license: mit
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
tags:
- llama
- lora
- political-science
- survey-replication
- canadian-election-study
- peft
- unsloth
datasets:
- custom
language:
- en
pipeline_tag: text-generation
---

# CES Phase 3B LoRA: With Party Identification

A LoRA adapter for Llama 3.1 8B Instruct that predicts political ideology using party identification in addition to leader ratings and policy positions.

**For most use cases, prefer [Phase 3A](https://huggingface.co/baglecake/ces-phase3a-lora) instead** — this model exists to demonstrate that party ID is redundant.

## Model Description

This model was trained on the Canadian Election Study (CES) 2021 to predict self-reported ideology (0-10 left-right scale) from:

- **Demographics**: Age, gender, province, education, employment, religion, marital status, urban/rural, born in Canada
- **Leader Thermometers**: Ratings (0-100) of Justin Trudeau, Erin O'Toole, and Jagmeet Singh
- **Wedge Issues**: Positions on carbon tax, energy/pipelines, and medical assistance in dying (MAID)
- **Government Satisfaction**: Overall satisfaction with federal government
- **Party Identification**: "I usually think of myself as a Liberal/Conservative/NDP..." (ONLY IN THIS MODEL)

## Performance

| Model | Inputs | Correlation (r) |
|-------|--------|-----------------|
| Phase 2 | Demographics + 3 psychographics | 0.428 |
| Phase 3A | + Leader thermometers + wedge issues | 0.560 |
| **Phase 3B (this model)** | **+ Party ID** | **0.574** |

**Partisan Delta = 0.014** — Party ID adds only 1.4% improvement.

## Why Phase 3A is Preferred

We trained this model (Phase 3B) specifically to test whether party identification adds predictive value beyond substantive attitudes. It doesn't.

**The null result is the finding:**
- Party identity is **redundant** — it's already encoded in how people feel about leaders and their policy positions
- Canadian ideology is **substantive, not tribal** — people's "team" reflects their actual views
- Adding party ID is "cheating" — you're just asking people their ideology with extra steps

Phase 3B exists for reproducibility and to demonstrate this null result empirically.

## Usage

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

base_model = AutoModelForCausalLM.from_pretrained(
    "meta-llama/Meta-Llama-3.1-8B-Instruct",
    load_in_4bit=True
)
model = PeftModel.from_pretrained(base_model, "baglecake/ces-phase3b-lora")
tokenizer = AutoTokenizer.from_pretrained("meta-llama/Meta-Llama-3.1-8B-Instruct")

# Example prompt (note: includes party ID)
system = """You are a 45-year-old man from Ontario, Canada. You live in a suburb of a large city. Your highest level of education is a bachelor's degree. You are currently employed full-time. You are married. You have children. You are Catholic. You were born in Canada.

Political Profile:
Leader Ratings: Justin Trudeau: 25/100, Erin O'Toole: 70/100, Jagmeet Singh: 30/100.
Views: Strongly disagrees that the federal government should continue the carbon tax; strongly agrees that the government should do more to help the energy sector/pipelines.
Overall Satisfaction: Is not at all satisfied with the federal government.
Party ID: Generally thinks of themselves as a Conservative.

Answer survey questions as this person would, based on their background and detailed political profile."""

user = "On a scale from 0 to 10, where 0 means left/liberal and 10 means right/conservative, where would you place yourself politically? Just give the number."
```

## Training Details

- **Base model**: meta-llama/Meta-Llama-3.1-8B-Instruct (4-bit quantized via Unsloth)
- **Training data**: 14,455 examples from CES 2021
- **LoRA rank**: 32
- **LoRA alpha**: 64
- **Target modules**: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj
- **Epochs**: 3
- **Hardware**: NVIDIA A100 40GB (Colab Pro)

## Limitations

1. **Narrow task**: Model only outputs ideology numbers (0-10).
2. **Canadian-specific**: Trained on CES 2021 under Trudeau government.
3. **Leader-specific**: Uses 2021 leader names.
4. **Redundant information**: Party ID doesn't add meaningful predictive value over Phase 3A.

## Citation

```bibtex
@software{ces-phase3-lora,
  title = {CES Phase 3 LoRA: Leader Affect and Policy Prediction},
  author = {Coburn, Del},
  year = {2025},
  url = {https://huggingface.co/baglecake/ces-phase3a-lora}
}
```

## Part of emile-GCE

This model is part of the [emile-GCE](https://github.com/delcoburn/emile-gce) project for Generative Computational Ethnography.