Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -18,6 +18,16 @@ size_categories:
|
|
| 18 |
|
| 19 |
Official dataset release of 4D steerability probe (reading difficulty, formality, textual diversity, text length goal-space). Initial probe contains 2,048 prompts used in our paper (32 different rewrites over 64 texts).
|
| 20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
## Results
|
| 22 |
|
| 23 |
Here, we will retain the most recent versions of SteerBench for top-performing models. Shown here: steering error of recent models (`mean (std. dev.)`).
|
|
@@ -101,5 +111,5 @@ Here, we will retain the most recent versions of SteerBench for top-performing m
|
|
| 101 |
<td>0.52 (0.22)</td>
|
| 102 |
</tr>
|
| 103 |
</table>
|
| 104 |
-
*
|
| 105 |
|
|
|
|
| 18 |
|
| 19 |
Official dataset release of 4D steerability probe (reading difficulty, formality, textual diversity, text length goal-space). Initial probe contains 2,048 prompts used in our paper (32 different rewrites over 64 texts).
|
| 20 |
|
| 21 |
+
## Dataset format
|
| 22 |
+
|
| 23 |
+
Each row contains a source text, along with its mappings in goal-space. We provide normalized and unnormalized values of the following for the source text:
|
| 24 |
+
* Flesch-Kincaid Grade Level (`reading_difficulty`)
|
| 25 |
+
* Heylighen-Dewaele F-Score (`formality`)
|
| 26 |
+
* Measure of Textual Lexical Diversity (`textual_diversity`)
|
| 27 |
+
* Word count (`text_length`)
|
| 28 |
+
|
| 29 |
+
We also provide goal vectors (`delta_*` or `target_*`) for all goal dimensions.
|
| 30 |
+
|
| 31 |
## Results
|
| 32 |
|
| 33 |
Here, we will retain the most recent versions of SteerBench for top-performing models. Shown here: steering error of recent models (`mean (std. dev.)`).
|
|
|
|
| 111 |
<td>0.52 (0.22)</td>
|
| 112 |
</tr>
|
| 113 |
</table>
|
| 114 |
+
* >1% invalid response rate
|
| 115 |
|