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

Feature #2416

Fires on “3” (not labeled yet: read the examples)

Fires on
1.12% of tokens
Mostly in
python
Layer
after block 4

Where it fires

  • python 55%
  • agent 20%
  • worksheet 13%
  • chat 8%
  • story 3%

Makes the model say

, picked again0 finish4 more pick

Makes it not say

))]))assert careople⏎⏎ love multipl

Strongest examples

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

  1. python 8.97
    , 7, 1]) == 4⏎assert evens([2, 7,
  2. python 8.94
    ]) == 4⏎assert max_item([3, 5,
  3. python 8.8
    returns the average of nums. For example, `mean_list([8, 1,
  4. python 8.68
    / len(nums)⏎```⏎⏎>>> average_list([3, 8,
  5. python 8.68
    returns the average of nums. For example, `mean_list([2, 6,
  6. python 8.65
    returns the number of items. For example, `count_items([3, 8,
  7. python 8.61
    returns the average of nums. For example, `average_list([8, 9,
  8. python 8.56
    , 5]) == [8]⏎assert keep_even([3, 6,
  9. python 8.52
    returns the average of nums. For example, `mean_list([3, 8,
  10. python 8.49
    / len(nums)⏎```⏎⏎>>> average_list([9, 1,
  11. python 8.47
    / len(nums)⏎```⏎⏎>>> mean_list([2, 4,
  12. python 8.44
    return result⏎```⏎⏎>>> sum_list([6, 9, 1])