Feature #2332
Fires on “nums” (not labeled yet: read the examples)
- Fires on
- 1.19% of tokens
- Mostly in
- agent
- Layer
- after block 4
Where it fires
- agent 52%
- python 30%
- story 11%
- worksheet 6%
- chat 2%
Makes the model say
sopath a trained decided i time
Makes it not say
They They she example girl SheShe high
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
agent 2.47functions: `cubed(x)` and `max_of(a, b)`.<|assistant_end|><|bos|>
-
agent 2.41result = result + n⏎13
-
python 2.41result = result + n⏎ return result
-
agent 2.29a coding agent harness.<|system_end|><|user_start|>Write a function only_even(nums) in core.
-
story 2.25to hurt Anna. He likes her. He says, "Wait, Anna. Come
-
story 2.24She was very excited.⏎Lucy's mom said, "Wait, we need to
-
python 2.24safe during the flood.<|bos|>`only_even(nums)` keeps only
-
python 2.23.py`, line 1.<|assistant_end|><|bos|>`keep_even(nums)` keeps only
-
story 2.22out of the box and followed her. It said, "Wait, Lily. Don
-
agent 2.22`add_one(x)` and `min_of(a, b)`.<|assistant_end|><|bos|>
-
agent 2.2functions: `average_list(nums)` and `doubled(nums)`.<|assistant_end|><|user_start|>re
-
agent 2.2` with `max_of(a, b)`.<|assistant_end|><|bos|>