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

Feature #606

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

Fires on
0.47% of tokens
Mostly in
python
Layer
after block 4

Where it fires

  • python 28%
  • agent 26%
  • chat 22%
  • story 12%
  • worksheet 10%

Makes the model say

:// exclamation calculator storyONuvestxt

Makes it not say

to few but# thanked back⏎$

Strongest examples

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

  1. worksheet 3.04
    03 more stickers. How many stickers does Max have
  2. chat 2.95
    Mia has 8 stickers. Mia gets 2 more stickers. How many stickers does Mia have
  3. chat 2.91
    Sue has 3 cookies. Sue gets 4 more cookies. How many cookies does Sue have
  4. chat 2.88
    has 74 cookies. Tim gets 0 more cookies. How many cookies does Tim have
  5. story 2.86
    that now they could play together with the yarn. The moral of the story is that sometimes
  6. chat 2.85
    10 toys. Ben gets 80 more toys. How many toys does Ben have
  7. chat 2.83
    Mia has 8 stickers. Mia gets 2 more stickers. How many stickers does Mia have
  8. story 2.81
    its life. They were happy and stayed friends forever. The moral of the story is: be
  9. chat 2.81
    Sam has 5 balls. Sam gets 2 more balls. How many balls does Sam have
  10. chat 2.8
    Sue has 8 toys. Sue gets 5 more toys. How many toys does Sue have
  11. chat 2.79
    42 balls. Sue gets 18 more balls. How many balls does Sue have
  12. story 2.76
    ed to the garden.⏎The moral of this story is that when