Skip to content
mini-lab
Features of prelude-2

Feature #2827

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

Fires on
0.76% of tokens
Mostly in
agent
Layer
after block 4

Where it fires

  • agent 84%
  • story 10%
  • chat 5%
  • worksheet 1%
  • python 0%

Makes the model say

they They" I he⏎>>> He

Makes it not say

nums reverventekMIT edit+ Instead

Strongest examples

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

  1. story 4.77
    a turtle swimming towards them.⏎The turtle said: "Don't you
  2. story 4.47
    fun.⏎But then, Sara feels tired. She says to Tom: "I want to
  3. agent 4.35
    aan(12, 2))⏎ ^^^^^⏎NameError: name 'mea
  4. agent 4.23
    print(posiitive(COUNT))⏎ ^^^^^^^^^⏎NameError: name 'pos
  5. agent 4.21
    be `positive`. It now prints:⏎```⏎True
  6. agent 4.18
    be `reverse`. It now prints:⏎```⏎r
  7. agent 4.18
    print(keep_veen(NUMBERS))⏎ ^^^^^^^^^⏎NameError: name 'keep_
  8. agent 4.16
    tiply(6, 1))⏎ ^^^^^^^^^⏎NameError: name 'mull
  9. agent 4.16
    , 8, 8, 0]))⏎ ^^^^^^^^⏎NameError: name 'mean_
  10. story 4.1
    go there anymore!"⏎Lily was very sad. She said to Jack: "Being popular
  11. agent 4.09
    7, 4, 3]))⏎ ^^^^^⏎NameError: name 'ta
  12. agent 4.09
    ⏎ print(niitial(WORD))⏎ ^^^^^^^⏎NameError: name 'ni