add results
Browse files
README.md
CHANGED
|
@@ -516,6 +516,43 @@ This repo includes scripts needed to run our full pipeline, including data prepr
|
|
| 516 |
| Subjectivity | ThatiAR | 2 | 2,446 | 748 | 467 |
|
| 517 |
|
| 518 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 519 |
## File Format
|
| 520 |
|
| 521 |
Each JSONL file in the dataset follows a structured format with the following fields:
|
|
|
|
| 516 |
| Subjectivity | ThatiAR | 2 | 2,446 | 748 | 467 |
|
| 517 |
|
| 518 |
|
| 519 |
+
## Results
|
| 520 |
+
|
| 521 |
+
Below, we present the performance of **LlamaLens** in **Arabic** compared to existing SOTA (if available) and the Llama-Instruct baseline, The “Delta” column here is
|
| 522 |
+
calculated as **(LLamalens – SOTA)**.
|
| 523 |
+
|
| 524 |
+
---
|
| 525 |
+
|
| 526 |
+
| **Task** | **Dataset** | **Metric** | **SOTA** | **Llama-instruct** | **LLamalens** | **Delta** (LLamalens - SOTA) |
|
| 527 |
+
|------------------------|---------------------------|-----------:|--------:|--------------------:|--------------:|------------------------------:|
|
| 528 |
+
| News Summarization | xlsum | R-2 | 0.137 | 0.034 | 0.075 | -0.062 |
|
| 529 |
+
| News Genre | ASND | Ma-F1 | 0.770 | 0.587 | 0.938 | 0.168 |
|
| 530 |
+
| News Genre | SANADAkhbarona | Acc | 0.940 | 0.784 | 0.922 | -0.018 |
|
| 531 |
+
| News Genre | SANADAlArabiya | Acc | 0.974 | 0.893 | 0.986 | 0.012 |
|
| 532 |
+
| News Genre | SANADAlkhaleej | Acc | 0.986 | 0.865 | 0.967 | -0.019 |
|
| 533 |
+
| News Genre | UltimateDataset | Ma-F1 | 0.970 | 0.376 | 0.883 | -0.087 |
|
| 534 |
+
| News Credibility | NewsCredibility | Acc | 0.899 | 0.455 | 0.494 | -0.405 |
|
| 535 |
+
| Emotion | Emotional-Tone | W-F1 | 0.658 | 0.358 | 0.748 | 0.090 |
|
| 536 |
+
| Emotion | NewsHeadline | Acc | 1.000 | 0.406 | 0.551 | -0.449 |
|
| 537 |
+
| Sarcasm | ArSarcasm-v2 | F1_Pos | 0.584 | 0.477 | 0.307 | -0.277 |
|
| 538 |
+
| Sentiment | ar_reviews_100k | F1_Pos | – | 0.343 | 0.665 | – |
|
| 539 |
+
| Sentiment | ArSAS | Acc | 0.920 | 0.603 | 0.795 | -0.125 |
|
| 540 |
+
| Stance | stance | Ma-F1 | 0.767 | 0.608 | 0.936 | 0.169 |
|
| 541 |
+
| Stance | Mawqif-Arabic-Stance | Ma-F1 | 0.789 | 0.764 | 0.867 | 0.078 |
|
| 542 |
+
| Att.worthiness | CT22Attentionworthy | W-F1 | 0.412 | 0.158 | 0.544 | 0.132 |
|
| 543 |
+
| Checkworthiness | CT24_T1 | F1_Pos | 0.569 | 0.404 | 0.877 | 0.308 |
|
| 544 |
+
| Claim | CT22Claim | Acc | 0.703 | 0.581 | 0.778 | 0.075 |
|
| 545 |
+
| Factuality | Arafacts | Mi-F1 | 0.850 | 0.210 | 0.534 | -0.316 |
|
| 546 |
+
| Factuality | COVID19Factuality | W-F1 | 0.831 | 0.492 | 0.781 | -0.050 |
|
| 547 |
+
| Propaganda | ArPro | Mi-F1 | 0.767 | 0.597 | 0.762 | -0.005 |
|
| 548 |
+
| Cyberbullying | ArCyc_CB | Acc | 0.863 | 0.766 | 0.753 | -0.110 |
|
| 549 |
+
| Harmfulness | CT22Harmful | F1_Pos | 0.557 | 0.507 | 0.508 | -0.049 |
|
| 550 |
+
| Hate Speech | annotated-hatetweets-4 | W-F1 | 0.630 | 0.257 | 0.549 | -0.081 |
|
| 551 |
+
| Hate Speech | OSACT4SubtaskB | Mi-F1 | 0.950 | 0.819 | 0.802 | -0.148 |
|
| 552 |
+
| Offensive | ArCyc_OFF | Ma-F1 | 0.878 | 0.489 | 0.652 | -0.226 |
|
| 553 |
+
| Offensive | OSACT4SubtaskA | Ma-F1 | 0.905 | 0.782 | 0.899 | -0.006 |
|
| 554 |
+
|
| 555 |
+
|
| 556 |
## File Format
|
| 557 |
|
| 558 |
Each JSONL file in the dataset follows a structured format with the following fields:
|