Feature #1963
Fires on “⏎” (not labeled yet: read the examples)
- Fires on
- 0.43% of tokens
- Mostly in
- python
- Layer
- after block 4
Where it fires
- python 65%
- agent 29%
- story 4%
- chat 2%
- worksheet 1%
Makes the model say
#$(- elementine+river checks
Makes it not say
she She hepy out here buy over
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
agent 3.85"⏎ ^^^^^^^^^^^^^^^^^^^^^^^^^^^⏎AssertionError⏎⏎Exited with code
-
agent 3.82, 4) == 16⏎ ^^^^^^^^^^^^^^^^^⏎AssertionError⏎⏎Exited with code
-
agent 3.815, 3]) == 9⏎ ^^^^^^^^^^^^^^^^^^^^^^^^^^^⏎AssertionError⏎⏎Exited with code
-
agent 3.777) == 1⏎ ^^^^^^^^^^^^^^^^⏎AssertionError⏎⏎Exited with code
-
python 3.73⏎⏎⏎def quiet(s):⏎ return s.lower()⏎⏎# stats.py
-
python 3.65(half(NUMBER))⏎print(odd(5))⏎⏎$ python3 app
-
agent 3.65== -12⏎ ^^^^^^^^^^^^^^^^^^^⏎AssertionError⏎⏎Exited with code
-
python 3.63+ 1⏎⏎⏎print(add_one(5))⏎⏎$ python3 main
-
agent 3.6217, 4) == False⏎ ^^^^^^^^^^^^^^^^^^^^^^^⏎AssertionError⏎⏎Exited with code
-
python 3.612⏎⏎⏎print(mean(10, 9))⏎⏎$ python3 main
-
agent 3.5914, 18]⏎ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^⏎AssertionError⏎⏎Exited with code
-
python 3.58, 4, 5, 1, 3]))⏎⏎$ python3 main