Feature #1837
Fires on “ example” (not labeled yet: read the examples)
- Fires on
- 0.36% of tokens
- Mostly in
- python
- Layer
- after block 4
Where it fires
- python 87%
- agent 7%
- story 3%
- worksheet 2%
- chat 1%
Makes the model say
again_*.it real example python okay.
Makes it not say
Y ate5iesotall jumping watch
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
python 4.02⏎```⏎⏎`subtract` subtracts b from a. For example, `subtract(
-
python 3.97b⏎```⏎⏎`subtract` returns a minus b. For example, `subtract(
-
python 3.85factorial` multiplies all the numbers from 1 to n. For example, `factorial(
-
python 3.841⏎```⏎⏎`previous` returns x minus one. For example, `previous(
-
python 3.812⏎```⏎⏎`double` returns twice x. For example, `double(
-
python 3.8```⏎⏎`excite` returns s with an exclamation mark. For example, `excite("
-
python 3.79⏎```⏎⏎`opposite` returns the opposite of x. For example, `opposite(
-
python 3.78⏎⏎`num_chars` returns the length of s. For example, `num_
-
python 3.78⏎⏎`mean` returns the average of a and b. For example, `mean(
-
python 3.74⏎⏎`num_chars` returns the length of s. For example, `num_
-
python 3.73```⏎⏎`upper_case` makes s upper case. For example, `upper_
-
python 3.71⏎⏎`bigger` returns the bigger of two numbers. For example, `bigger