Skip to content
mini-lab
Features of prelude-2

Feature #3962

Fires on โ€œ,โ€ (not labeled yet: read the examples)

Fires on
0.26% of tokens
Mostly in
story
Layer
after block 4

Where it fires

  • story 75%
  • chat 23%
  • agent 2%
  • python 0%
  • worksheet 0%

Makes the model say

and thereakesโŽ forg or Tim can

Makes it not say

]path<|arg|>]`.Once feeling functionsur

Strongest examples

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

  1. story 2.82
    Add is_divisible to ops.py<|assistant_end|><|bos|>Once upon a time, in a small town
  2. story 2.78
    only tell short stories and add numbers.<|assistant_end|><|bos|>Once upon a time, in a quiet little
  3. story 2.78
    can tell short stories and add numbers.<|assistant_end|><|bos|>Once upon a time, there was a little
  4. story 2.76
    how to tell stories and add numbers.<|assistant_end|><|bos|>Once upon a time, there was a little
  5. story 2.74
    only tell short stories and add numbers.<|assistant_end|><|bos|>Once upon a time, there was a big
  6. story 2.72
    can tell short stories and add numbers.<|assistant_end|><|bos|>Once upon a time, there was a naughty
  7. story 2.71
    json` and `demo.py`.<|assistant_end|><|bos|>Once upon a time, there was a girl
  8. story 2.7
    and `test_arith.py`.<|assistant_end|><|bos|>Once upon a time, there was a little
  9. story 2.7
    s<|user_end|><|assistant_start|>Anytime!<|assistant_end|><|bos|>Once upon a time, there was a little
  10. story 2.69
    how to tell stories and add numbers.<|assistant_end|><|bos|>Once upon a time, there was a monkey
  11. story 2.68
    2=9, 9<|think_end|>9<|assistant_end|><|bos|>Once upon a time, in a small messy
  12. story 2.68
    . Try asking me for a story!<|assistant_end|><|bos|>Once upon a time, there was a fat