Skip to content
mini-lab
Log in Sign up
Features of prelude-2

Feature #1833

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

Fires on
0.17% of tokens
Mostly in
story
Layer
after block 4

Where it fires

  • story 57%
  • agent 21%
  • worksheet 13%
  • chat 8%
  • python 1%

Makes the model say

,. mean or bad went: out

Makes it not say

countscalculator finished`: prelude Soon functionre

Strongest examples

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

  1. agent 1.66
    upper_ase' is not defined. Did you mean: 'upper_
  2. agent 1.54
    'lenggth' is not defined. Did you mean: 'length'?
  3. agent 1.52
    'totaal' is not defined. Did you mean: 'total'?
  4. story 1.48
    are my eggs." Lily and Tom were very surprised. They did not know gems were bird
  5. agent 1.47
    'double_al' is not defined. Did you mean: 'double_
  6. story 1.44
    made a big mess and broke the musician's hat. The musician
  7. story 1.43
    should never touch knives. They are very dangerous. They can hurt you badly."
  8. story 1.42
    serve them."⏎But when they opened the prune box, they found a little kitten
  9. story 1.4
    was very happy to help.⏎But when Tim opened the fridge, he saw a big cake
  10. story 1.38
    a gurgling noise from the north. He guessed that it must have
  11. story 1.38
    refrigerator seemed so big. But as soon as she opened the door, she saw that the fridge
  12. story 1.37
    . You can't play with it, it's dangerous. You could hurt yourself or someone else