Feature #2558
Fires on “<|system_end|>” (not labeled yet: read the examples)
- Fires on
- 0.48% of tokens
- Mostly in
- agent
- Layer
- after block 4
Where it fires
- agent 83%
- chat 16%
- story 1%
- worksheet 0%
- python 0%
Makes the model say
replaceAll >ily glad continued sorryem cheered
Makes it not say
little ` big y but that p first
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
agent 3.49ch, websearch, write, execute<|system_end|><|system_start|>You are an
-
agent 3.46ch, websearch, write, execute<|system_end|><|system_start|>You are an
-
agent 3.45ch, websearch, write, execute<|system_end|><|system_start|>You are an
-
agent 3.44ch, websearch, write, execute<|system_end|><|system_start|>You are an
-
agent 3.43ch, websearch, write, execute<|system_end|><|system_start|>You are an
-
agent 3.43ch, websearch, write, execute<|system_end|><|system_start|>You are an
-
agent 3.42ch, websearch, write, execute<|system_end|><|system_start|>You are an
-
agent 3.42ch, websearch, write, execute<|system_end|><|system_start|>You are an
-
agent 3.42ch, websearch, write, execute<|system_end|><|system_start|>You are an
-
agent 3.41ch, websearch, write, execute<|system_end|><|system_start|>You are an
-
agent 3.4ch, websearch, write, execute<|system_end|><|system_start|>You are an
-
agent 3.4ch, websearch, write, execute<|system_end|><|system_start|>You are an