Add Results
Browse files
README.md
CHANGED
|
@@ -330,6 +330,32 @@ This repo includes scripts needed to run our full pipeline, including data prepr
|
|
| 330 |
| Subjectivity | clef2024-checkthat-lab | 2 | 825 | 484 | 219 |
|
| 331 |
|
| 332 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 333 |
## File Format
|
| 334 |
|
| 335 |
Each JSONL file in the dataset follows a structured format with the following fields:
|
|
|
|
| 330 |
| Subjectivity | clef2024-checkthat-lab | 2 | 825 | 484 | 219 |
|
| 331 |
|
| 332 |
|
| 333 |
+
## Results
|
| 334 |
+
|
| 335 |
+
Below, we present the performance of **LlamaLens** in **English** compared to existing SOTA (if available) and the Llama-Instruct baseline, The “Delta” column here is
|
| 336 |
+
calculated as **(LLamalens – SOTA)**.
|
| 337 |
+
|
| 338 |
+
| **Task** | **Dataset** | **Metric** | **SOTA** | **Llama-instruct** | **LLamalens** | **Delta** (LLamalens - SOTA) |
|
| 339 |
+
|----------------------|---------------------------|-----------:|--------:|--------------------:|--------------:|------------------------------:|
|
| 340 |
+
| News Summarization | xlsum | R-2 | 0.152 | 0.074 | 0.141 | -0.011 |
|
| 341 |
+
| News Genre | CNN_News_Articles | Acc | 0.940 | 0.644 | 0.915 | -0.025 |
|
| 342 |
+
| News Genre | News_Category | Ma-F1 | 0.769 | 0.970 | 0.505 | -0.264 |
|
| 343 |
+
| News Genre | SemEval23T3-ST1 | Mi-F1 | 0.815 | 0.687 | 0.241 | -0.574 |
|
| 344 |
+
| Subjectivity | CT24_T2 | Ma-F1 | 0.744 | 0.535 | 0.508 | -0.236 |
|
| 345 |
+
| Emotion | emotion | Ma-F1 | 0.790 | 0.353 | 0.878 | 0.088 |
|
| 346 |
+
| Sarcasm | News-Headlines | Acc | 0.897 | 0.668 | 0.956 | 0.059 |
|
| 347 |
+
| Sentiment | NewsMTSC | Ma-F1 | 0.817 | 0.628 | 0.627 | -0.190 |
|
| 348 |
+
| Checkworthiness | CT24_T1 | F1_Pos | 0.753 | 0.404 | 0.877 | 0.124 |
|
| 349 |
+
| Claim | claim-detection | Mi-F1 | – | 0.545 | 0.915 | – |
|
| 350 |
+
| Factuality | News_dataset | Acc | 0.920 | 0.654 | 0.946 | 0.026 |
|
| 351 |
+
| Factuality | Politifact | W-F1 | 0.490 | 0.121 | 0.290 | -0.200 |
|
| 352 |
+
| Propaganda | QProp | Ma-F1 | 0.667 | 0.759 | 0.851 | 0.184 |
|
| 353 |
+
| Cyberbullying | Cyberbullying | Acc | 0.907 | 0.175 | 0.847 | -0.060 |
|
| 354 |
+
| Offensive | Offensive_Hateful | Mi-F1 | – | 0.692 | 0.805 | – |
|
| 355 |
+
| Offensive | offensive_language | Mi-F1 | 0.994 | 0.646 | 0.884 | -0.110 |
|
| 356 |
+
| Offensive & Hate | hate-offensive-speech | Acc | 0.945 | 0.602 | 0.924 | -0.021 |
|
| 357 |
+
|
| 358 |
+
|
| 359 |
## File Format
|
| 360 |
|
| 361 |
Each JSONL file in the dataset follows a structured format with the following fields:
|