Feature #3917
Fires on “read” (not labeled yet: read the examples)
- Fires on
- 0.75% of tokens
- Mostly in
- agent
- Layer
- after block 4
Where it fires
- agent 56%
- chat 20%
- python 14%
- story 5%
- worksheet 4%
Makes the model say
<|arg|> table lookedub near watch functions
Makes it not say
⏎tet":red):<|bos|>
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
agent 1.63.py. It returns the first letter of s.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=lib
-
agent 1.57stats.py: it returns the average of nums.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=stats
-
agent 1.53to greetings.py: it makes s upper case.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=greetings
-
agent 1.52<|user_start|>Please which functions are in strings.py?<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=strings
-
agent 1.51py. It returns the average of a and b.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=utils
-
agent 1.51, what does negate return?<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=ops
-
agent 1.51<|system_end|><|user_start|>explain product from number_utils.py<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=number
-
agent 1.5<|user_start|>What's in items.py?<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=items
-
agent 1.49py: it returns the sum of a and b.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=math
-
chat 1.4904 + 58?<|user_end|><|assistant_start|><|tool_call_start|>calculator<|arg|>expression=204 +
-
story 1.48Mia says.⏎Tom shakes his head. He knows the truth. Flow
-
chat 1.483 and 3 for me?<|user_end|><|assistant_start|><|tool_call_start|>calculator<|arg|>expression=53 + 3