Skip to content
mini-lab
Features of prelude-2

Feature #1179

Fires on “1” (not labeled yet: read the examples)

Fires on
1.65% of tokens
Mostly in
worksheet
Layer
after block 4

Where it fires

  • worksheet 48%
  • chat 38%
  • agent 13%
  • story 1%
  • python 1%

Makes the model say

23015<|assistant_end|>erell

Makes it not say

what there the tailFixed lotspattern

Strongest examples

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

  1. chat 2.9
    1=11, 113<|think_end|>The answer is 113.<|assistant_end|>
  2. chat 2.88
    +0=1, 140<|think_end|>The answer is 140.<|assistant_end|>
  3. chat 2.87
    10, 1002<|think_end|>The answer is 1002.<|assistant_end|>
  4. chat 2.83
    10, 1006<|think_end|>The answer is 1006.<|assistant_end|>
  5. chat 2.83
    =13, 133<|think_end|>The answer is 133.<|assistant_end|><|bos|>
  6. chat 2.83
    12, 1248<|think_end|>The answer is 1248.<|assistant_end|>
  7. chat 2.81
    +7=7, 766<|think_end|>The answer is 766.<|assistant_end|>
  8. chat 2.79
    8=17, 172<|think_end|>The answer is 172.<|assistant_end|>
  9. chat 2.78
    3=10, 10<|think_end|>The answer is 10.<|assistant_end|><|bos|><|user_start|>
  10. chat 2.78
    5<|tool_call_end|><|assistant_end|><|tool_start|>137<|tool_end|><|assistant_start|>The answer is 137.<|assistant_end|><|user_start|>
  11. chat 2.78
    =10, 100<|think_end|>The answer is 100.<|assistant_end|><|bos|>
  12. chat 2.77
    +9=15, 15<|think_end|>The answer is 15.<|assistant_end|><|bos|>