Feature #45
Fires on “ in” (not labeled yet: read the examples)
- Fires on
- 0.25% of tokens
- Mostly in
- story
- Layer
- after block 4
Where it fires
- story 53%
- agent 31%
- chat 8%
- python 6%
- worksheet 2%
Makes the model say
!" before!', of," morning
Makes it not say
cat cross boy Tim Lily]`. himjson
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
story 3.32on it. It was
-
chat 3.32on, they had many
-
story 3.32on the slide too.
-
story 3.32on a folder. "
-
story 3.32on his hind legs
-
python 3.1in nums]⏎⏎⏎
-
agent 3.1in `lib.py
-
story 3.1in the grass. They
-
story 3.1in his house.⏎
-
python 3.1in nums]` with `
-
agent 3.1in OpenCode, a coding
-
agent 3.1in `items.py