Feature #2516
Fires on “ liked” (not labeled yet: read the examples)
- Fires on
- 0.45% of tokens
- Mostly in
- story
- Layer
- after block 4
Where it fires
- story 72%
- chat 12%
- python 7%
- agent 6%
- worksheet 3%
Makes the model say
i toake playing editourython[
Makes it not say
equals,"⏎!#?"MyThe
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
story 6.36friends to play with every day.<|bos|>Ben is a big boy who likes to help his mom
-
story 6.32was always there to protect him.<|bos|>Tom is a big dog who likes to play with his
-
story 5.87houses, happy with their tasty snack.<|bos|>Ben is a boy who likes to dig in the
-
story 5.85and add numbers.<|assistant_end|><|bos|>Lila and Sam were best friends. They liked to play in the
-
story 5.81and add numbers.<|assistant_end|><|bos|>Anna and Lily were best friends. They liked to play with dolls
-
story 5.77tell short stories and add numbers.<|assistant_end|><|bos|>Tom is a boy who likes to play with his
-
story 5.77scared of the water anymore.<|assistant_end|><|bos|>Lila was a tall girl who liked to play outside.
-
story 5.65is 552.<|assistant_end|><|bos|>Lily was a green bird who liked to talk a lot
-
story 5.64333.<|assistant_end|><|bos|>Ben and Sam were brothers who liked to play with blocks
-
story 5.546])⏎17⏎<|bos|>Lily and Anna were sisters who liked to play dress-
-
story 5.52and add numbers.<|assistant_end|><|bos|>Lily and Ben were best friends. They liked to play in the
-
story 5.516<|think_end|>6<|assistant_end|><|bos|>Lily and Tom were best friends. They liked to play in the