Skip to content
mini-lab
Features of prelude-2

Feature #1281

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

Fires on
3.37% of tokens
Mostly in
agent
Layer
after block 4

Where it fires

  • agent 38%
  • worksheet 28%
  • chat 22%
  • python 11%
  • story 1%

Makes the model say

102 atOnce (Yes

Makes it not say

, Lily one walked there up way as

Strongest examples

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

  1. story 6.23
    played with the toy all day. The other dog saw this and felt envious too.
  2. chat 6.08
    <|user_start|>Tell me a short story.<|user_end|><|assistant_start|>Bob was feeling excited. He was going to
  3. chat 5.72
    in the park. He was jumping from
  4. story 5.7
    big dog named Max. Max was very fast and strong. Tom felt envious of Max
  5. story 5.66
    on the stay. Flip was very worried and regretful. Quickly
  6. chat 5.65
    was a grumpy boat. The boat was always
  7. story 5.63
    his friend, Sue, eating a big ice cream cone. Tim felt envious. He
  8. story 5.53
    lunch. Whenever he did this, Jack would be so envious that he
  9. chat 5.36
    a tree. He looked at his
  10. python 5.27
    python⏎def sum_up_to(n):⏎ acc = 0⏎ for i
  11. python 5.18
    ⏎⏎⏎def count_vowels(s):⏎ k = 0⏎ for c
  12. story 5.17
    , the dog saw another dog with a big toy. The dog felt envious. He