Feature #21
Fires on “:” (not labeled yet: read the examples)
- Fires on
- 0.42% of tokens
- Mostly in
- agent
- Layer
- after block 4
Where it fires
- agent 67%
- python 10%
- chat 10%
- story 8%
- worksheet 5%
Makes the model say
happy learnedre loved head love lastE
Makes it not say
=} % So⏎ uponexpressioncommand
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
python 3.04sum_list` should add up all the numbers in nums
-
python 2.96⏎ return result⏎```⏎⏎`sum_list` adds up all the numbers in nums
-
agent 2.88all tests passed⏎<|tool_end|><|assistant_start|>`test_words.py` passes: all tests passed.
-
agent 2.87all tests passed⏎<|tool_end|><|assistant_start|>`test_greetings.py` passes: all tests passed.
-
agent 2.86all tests passed⏎<|tool_end|><|assistant_start|>`test_utils.py` passes: all tests passed.
-
agent 2.84all tests passed⏎<|tool_end|><|assistant_start|>`test_words.py` passes: all tests passed.
-
agent 2.8all tests passed⏎<|tool_end|><|assistant_start|>`test_words.py` passes: all tests passed.
-
agent 2.79all tests passed⏎<|tool_end|><|assistant_start|>`test_items.py` passes: all tests passed.
-
agent 2.79all tests passed⏎<|tool_end|><|assistant_start|>`test_lists.py` passes: all tests passed.
-
agent 2.79all tests passed⏎<|tool_end|><|assistant_start|>`test_words.py` passes: all tests passed.
-
agent 2.78all tests passed⏎<|tool_end|><|assistant_start|>`test_utils.py` passes: all tests passed.
-
agent 2.77all tests passed⏎<|tool_end|><|assistant_start|>`test_tools.py` passes: all tests passed.