Feature #1886
Fires on “NUMBERS” (not labeled yet: read the examples)
- Fires on
- 0.59% of tokens
- Mostly in
- agent
- Layer
- after block 4
Where it fires
- agent 67%
- python 19%
- story 9%
- chat 4%
- worksheet 2%
Makes the model say
= plusCanus somethingive`in
Makes it not say
upon okayFixedaid anyone parkls They
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
python 6.49py⏎from math_utils import larger, sub_one⏎⏎COUNT = 6⏎
-
python 6.49py⏎from number_utils import magnitude, lower_of⏎⏎NUMBER = -9⏎
-
python 6.44⏎from listlib import evens, count⏎from ops import minus⏎⏎NUMBERS = [5,
-
python 6.16of, decrement⏎from number_utils import abs_value⏎⏎SIZE = 1⏎
-
python 6.14words import combine, backwards⏎from items import double_all⏎⏎NUMBERS = [0,
-
python 6.09py⏎from mathlib import previous⏎from stats import head⏎⏎SCORES = [9,
-
python 6.05py⏎from math_utils import subtract, factorial, halve⏎⏎SIZE = 11
-
python 5.99.<|assistant_end|><|bos|># main.py⏎from items import biggest⏎⏎VALUES = [1,
-
python 5.98demo.py⏎from listlib import total, min_item⏎⏎NUMBERS = [9,
-
python 5.89py⏎from stats import mean_list, keep_even⏎⏎NUMBERS = [5,
-
python 5.84(x):⏎ return x % 2 == 1⏎⏎⏎NUMBER = 10
-
python 5.79run.py⏎from items import head, double_all⏎⏎VALUES = [8,