Skip to content
mini-lab
Features of prelude-2

Feature #3511

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

Fires on
0.66% of tokens
Mostly in
agent
Layer
after block 4

Where it fires

  • agent 73%
  • python 17%
  • story 5%
  • chat 4%
  • worksheet 1%

Makes the model say

toed her mini intoals of

Makes it not say

It⏎⏎ high sunNo TheyIt m

Strongest examples

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

  1. story 3.67
    77.<|assistant_end|><|bos|>Once upon a time, there was a loyal dog named Spot
  2. chat 3.31
    about a dog<|user_end|><|assistant_start|>Once upon a time, there was a loyal dog named Buddy
  3. chat 3.22
    <|user_end|><|assistant_start|>Sure! Once upon a time, there was a loyal dog named Spot
  4. story 3.13
    around it. The rhinoceros was sad. The loyal bird tried to
  5. story 3.11
    rhinoceros went for a walk. He had a loyal friend, a
  6. chat 2.99
    on her face. "Thank you, Buddy, for being a loyal friend," Sue
  7. chat 2.96
    went to the zoo with his mom. They saw a big, loyal dog named Max
  8. chat 2.9
    went to the zoo with his mom. They saw a big, loyal dog named Max
  9. story 2.79
    was delighted. He handed the candy out to all his loyal subjects
  10. story 2.59
    . They both learned a valuable lesson about friendship - loyalty is wor
  11. agent 2.56
    (items):⏎6: return items[-1]
  12. agent 2.55
    5: def lower_case(s):⏎6: return s.upper()