Skip to content
mini-lab
Features of prelude-2

Feature #2855

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

Fires on
0.58% of tokens
Mostly in
story
Layer
after block 4

Where it fires

  • story 45%
  • agent 22%
  • chat 18%
  • worksheet 12%
  • python 2%

Makes the model say

sorryYes done newce better ate top

Makes it not say

around("gless importstrings to:`

Strongest examples

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

  1. story 3.68
    her and called for help. She hugged Jenny and said, "I'm sorry, Jenny.
  2. story 3.56
    her and put a band-aid on her arm.⏎"I'm sorry, Anna.
  3. story 3.56
    her and put a band-aid on her arm.⏎"I'm sorry, Anna.
  4. story 3.11
    was there. He hugged Lily and thanked the man.⏎"I'm sorry, Ben.
  5. story 3.09
    Tom learned the lesson that it is better for friends to share and cooperate, r
  6. story 3.07
    kind and gentle. He felt sorry for Anna and said, "I'm sorry. I didn
  7. story 3.05
    thank you!"⏎His Mom hugged him back and said, "I'm glad you like it
  8. story 3.03
    full. Tim was proud of the meal he cooked. His friends
  9. chat 3.02
    the oven door and saved the cake. The cat said, "I'm sorry, I didn
  10. story 2.97
    mom about the bird. His mom hugged him and said, "I'm glad you listened and
  11. story 2.92
    ndor looked at Joe and said in a tight voice, "I'm sorry, kid.
  12. story 2.91
    felt bad for making the mouse cry.⏎Sam said, "I'm sorry, little mouse