Research map / Large language models / Language, safety and interpretability
Predicting and Analyzing Language Specificity in Social Media Posts
Yifan Gao, Yang Zhong, Daniel Preoţiuc-Pietro, Junyi Jessy Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add surface & lexical features (ablation) | all-features SVR · Tweet specificity test set | MAE: 0.3856 -> 0.3583 | Improved |
| add distributional word representations (ablation) | all-features SVR · Tweet specificity test set | MAE: 0.3965 -> 0.3583 | Improved |
| add emotion features (ablation) | all-features SVR · Tweet specificity test set | MAE: 0.3617 -> 0.3583 | Improved |
| add social media content features (URL, reply) (ablation) | all-features SVR · Tweet specificity test set | MAE: 0.3578 -> 0.3583 | No clear effect |
| add social media features (ablation) | all-features SVR · Tweet specificity test set | Pearson correlation: 0.735 -> 0.734 | No clear effect |
| use trained feature model instead of average predictor (ablation) | SVR with all features · Tweet specificity test set | MAE: 0.5318 -> 0.3583 | Improved |
| use all features instead of length alone | SVR · Tweet specificity test set | MAE: 0.4726 -> 0.3583 | Improved |
| use all features instead of Speciteller trained on news | SVR · Tweet specificity test set | MAE: 1.0614 -> 0.3583 | Improved |
| use word representations alone vs best baseline (ablation) | SVR · Tweet specificity test set | MAE: 0.4726 -> 0.3885 | Improved |
| use social media content alone vs length baseline (ablation) | SVR · Tweet specificity test set | MAE: 0.4726 -> 0.5180 | Hurt |
| use emotion features alone vs average baseline (ablation) | SVR · Tweet specificity test set | MAE: 0.5318 -> 0.5324 | No clear effect |
| use all features vs word representations alone (ablation) | SVR · Tweet specificity test set | Pearson correlation: 0.661 -> 0.734 | Improved |
Closest papers by meaning
- ImpScore: A Learnable Metric For Quantifying The Implicitness Level of Sentences (ICLR 2025 · similarity 0.83)
- Fine-Tuning Language Models for Factuality (ICLR 2024 · similarity 0.83)
- Narrowing the Gap between Supervised and Unsupervised Sentence Representation Learning with Large Language Model (AAAI 2024 · similarity 0.82)
- Language Model Meets Prototypes: Towards Interpretable Text Classification Models through Prototypical Networks (AAAI 2025 · similarity 0.82)
- PersonalSum: A User-Subjective Guided Personalized Summarization Dataset for Large Language Models (NeurIPS 2024 · similarity 0.81)
- Paloma: A Benchmark for Evaluating Language Model Fit (NeurIPS 2024 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
