Datasets:
Card and manifest
Browse files
README.md
CHANGED
|
@@ -17,7 +17,7 @@ size_categories:
|
|
| 17 |
|
| 18 |
# dramatis v4 data
|
| 19 |
|
| 20 |
-
The data behind the paper *
|
| 21 |
models on controlled corpora of children's stories and asks whether a model builds a
|
| 22 |
"persona" for a class of character names. This repo holds the 1.27 million generated stories,
|
| 23 |
the probe and finetune sets, the recipes that turn the stories into each training corpus, and
|
|
@@ -113,10 +113,10 @@ models.
|
|
| 113 |
## Citation
|
| 114 |
|
| 115 |
```bibtex
|
| 116 |
-
@misc{
|
| 117 |
-
title
|
| 118 |
-
author
|
| 119 |
-
year
|
| 120 |
-
|
| 121 |
}
|
| 122 |
```
|
|
|
|
| 17 |
|
| 18 |
# dramatis v4 data
|
| 19 |
|
| 20 |
+
The data behind the paper *Toy emergent misalignment in small language models pretrained on controlled corpora* (Satzinger, 2026), which trains small language
|
| 21 |
models on controlled corpora of children's stories and asks whether a model builds a
|
| 22 |
"persona" for a class of character names. This repo holds the 1.27 million generated stories,
|
| 23 |
the probe and finetune sets, the recipes that turn the stories into each training corpus, and
|
|
|
|
| 113 |
## Citation
|
| 114 |
|
| 115 |
```bibtex
|
| 116 |
+
@misc{satzinger2026toy,
|
| 117 |
+
title = {Toy emergent misalignment in small language models pretrained on controlled corpora},
|
| 118 |
+
author = {Satzinger, Kevin J.},
|
| 119 |
+
year = {2026},
|
| 120 |
+
howpublished = {\url{https://github.com/kjsatz/dramatis}}
|
| 121 |
}
|
| 122 |
```
|