Skip to content
mini-lab
Features of prelude-2

Feature #837

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

Fires on
0.52% of tokens
Mostly in
chat
Layer
after block 4

Where it fires

  • chat 46%
  • agent 33%
  • story 15%
  • python 4%
  • worksheet 2%

Makes the model say

idea too % with for at nodded.

Makes it not say

list⏎ moralFixedrement Soonassertbox

Strongest examples

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

  1. chat 2.49
    jumping from tree to tree.<|assistant_end|><|user_start|>I want a story about a princess!<|user_end|><|assistant_start|>Once
  2. chat 2.48
    a helpful assistant.<|system_end|><|user_start|>Can you tell me a story about a princess?<|user_end|><|assistant_start|>Once
  3. chat 2.38
    is 823.<|assistant_end|><|user_start|>I want a story about a princess!<|user_end|><|assistant_start|>Once
  4. chat 2.35
    920.<|assistant_end|><|user_start|>Can you tell me a story about a princess?<|user_end|><|assistant_start|>Once
  5. chat 2.35
    23.<|assistant_end|><|user_start|>Tell me a story about a dragon.<|user_end|><|assistant_start|>Once
  6. chat 2.31
    Answer in one short sentence.<|system_end|><|user_start|>I want a story about a princess!<|user_end|><|assistant_start|>Once
  7. chat 2.3
    <|assistant_end|><|bos|><|system_start|>tools: calculator<|system_end|><|user_start|>Tell me a story about a dragon.<|user_end|><|assistant_start|>Once
  8. chat 2.29
    105.<|assistant_end|><|bos|><|user_start|>Write a short story about a princess.<|user_end|><|assistant_start|>Once
  9. chat 2.29
    Answer in one short sentence.<|system_end|><|user_start|>I want a story about a princess!<|user_end|><|assistant_start|>Once
  10. chat 2.25
    Answer in one short sentence.<|system_end|><|user_start|>I want a story about a princess!<|user_end|><|assistant_start|>Once
  11. chat 2.25
    every answer with "Sure!".<|system_end|><|user_start|>Write a short story about a princess.<|user_end|><|assistant_start|>Sure
  12. chat 2.24
    <|think_end|>509<|assistant_end|><|bos|><|system_start|>tools: calculator<|system_end|><|system_start|>Do not use the calculator.<|system_end|>