Skip to content
mini-lab
Features of prelude-2

Feature #2071

Fires on “ to” (not labeled yet: read the examples)

Fires on
0.37% of tokens
Mostly in
agent
Layer
after block 4

Where it fires

  • agent 33%
  • worksheet 29%
  • story 19%
  • chat 13%
  • python 6%

Makes the model say

9 is10 equals]) tree hill

Makes it not say

` " 'ideter "!"`. everyb

Strongest examples

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

  1. chat 2.09
    in this project.<|assistant_end|><|bos|><|user_start|>thanks<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|user_start|>
  2. chat 2.07
    to playing normally.<|bos|><|user_start|>thanks<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|user_start|>
  3. chat 2.05
    for another story soon.<|assistant_end|><|user_start|>thanks<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|user_start|>
  4. chat 2.04
    <|assistant_start|>You're welcome!<|assistant_end|><|user_start|>Thanks!<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|user_start|>
  5. chat 2.02
    5])⏎2⏎<|bos|><|user_start|>Thanks!<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|user_start|>
  6. chat 1.96
    !<|assistant_end|><|bos|><|user_start|>Great, thanks!<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|user_start|>
  7. chat 1.9
    assistant.<|system_end|><|user_start|>Great, thanks!<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|bos|>
  8. chat 1.89
    .<|assistant_end|><|bos|><|user_start|>Great, thanks!<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|user_start|>
  9. chat 1.88
    Have a nice day.<|assistant_end|><|user_start|>Thank you<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|user_start|>
  10. chat 1.87
    or with a calculator.<|assistant_end|><|user_start|>thanks<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|bos|>
  11. chat 1.86
    is 10.<|assistant_end|><|user_start|>thanks<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|user_start|>
  12. chat 1.85
    <|bos|><|system_start|>tools: calculator<|system_end|><|user_start|>thanks<|user_end|><|assistant_start|>Happy to help!<|assistant_end|><|user_start|>