Feature #4034
Fires on “newString” (not labeled yet: read the examples)
- Fires on
- 0.55% of tokens
- Mostly in
- agent
- Layer
- after block 4
Where it fires
- agent 89%
- chat 8%
- story 3%
- python 0%
- worksheet 0%
Makes the model say
= They onlyRename are about uponpered
Makes it not say
my5$lii and little1
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
agent 3.81path=utils.py<|arg|>oldString= return a * b<|arg|>newString= return a
-
agent 3.77.py<|arg|>oldString= return s.lower()<|arg|>newString= return s
-
agent 3.77test_arith.py<|arg|>oldString=positive<|arg|>newString=is_positive
-
agent 3.75= return s[-1]<|arg|>newString= return s
-
agent 3.75=strings.py<|arg|>oldString= return s.upper()<|arg|>newString= return s
-
agent 3.7=core.py<|arg|>oldString= return s.upper()<|arg|>newString= return s
-
agent 3.68path=core.py<|arg|>oldString= return a - b<|arg|>newString= return a
-
agent 3.64=test_calc.py<|arg|>oldString=absolute<|arg|>newString=magnitude<|arg|>
-
agent 3.63path=ops.py<|arg|>oldString= return a * b<|arg|>newString= return a
-
agent 3.63math_utils.py<|arg|>oldString= return a * b<|arg|>newString= return a
-
agent 3.63path=calc.py<|arg|>oldString= return a * b<|arg|>newString= return a
-
agent 3.62=helpers.py<|arg|>oldString= return s.upper()<|arg|>newString= return s