demo-social-media

demo-social-media

This model was trained using influence-guided dataset selection, a technique that uses influence scores to identify the most impactful training data for specific concepts.

Model Description

• Base Model: distilgpt2
• Training Concepts: sentiment, analysis, on, social, media
• Training Method: Influence-guided data selection
• Compute Budget: 100 steps per condition
• Total Datasets: 4

Training Approach

This model was trained using three different data selection strategies to validate the effectiveness of influence-guided training:

1. Positive Influence: Datasets with high positive influence scores (most aligned with target concepts)
2. Random Baseline: Randomly sampled datasets
3. Negative Influence: Datasets with high negative influence scores (least aligned)

Benchmark Results

Lower is better for all metrics

Training Datasets

The model was trained on datasets selected through influence scoring:

• google-research-datasets/poem_sentiment (Influence: 46.167)
• patched-codes/static-analysis-eval (Influence: -13.218)
• dvilasuero/ag_news_error_analysis (Influence: -16.302)
• MahdiA/Iran-protests-media (Influence: 90.503)

Intended Use

This model demonstrates the effectiveness of influence-guided training for:

• Concept-specific language modeling
• Data-efficient training
• Dataset curation research

Limitations

• Trained on a limited compute budget for benchmarking purposes
• May not generalize well outside the target concepts: sentiment, analysis, on, social, media
• Performance depends on the quality of influence score estimation

Citation

If you use this model or the influence-guided training approach, please cite:

@software{influence_guided_training,
  title = {Influence-Guided Dataset Selection for Language Models},
  author = {Dowser by Durinn},
  year = {2025},
  url = {https://huggingface.co/durinn/demo-social-media}
}

Model Card Contact

For questions or feedback, visit Durinn

---

Generated by Dowser - Dataset discovery and training optimization