Feature #24

Strongest examples

On text the dictionary never trained on. The underlined token is where it fires most.

1. agent 3.15
   .py. It returns the first letter of s.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=lib
2. agent 3.08
   .py that returns the first letter of s please"<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=utils
3. agent 3.01
   n), which repeats s n times.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=tools
4. agent 3.0
   reverse in text.py that returns s reversed.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=text
5. agent 2.99
   to textlib.py. It makes s upper case."<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=textlib
6. agent 2.98
   nums), which doubles every number in nums.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=utils
7. agent 2.97
   py. It returns the number of characters in s.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=strings
8. agent 2.96
   .py that returns the average of a and b.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=main
9. agent 2.95
   to greetings.py: it makes s upper case.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=greetings
10. agent 2.95
    3, 3] in demo.py.<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=demo
11. agent 2.91
    .py that returns the sum of the numbers in nums<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=utils
12. agent 2.91
    greetings.py that returns the first letter of s please<|user_end|><|assistant_start|><|tool_call_start|>read<|arg|>path=greetings