Feature #1340
Fires on “ running” (not labeled yet: read the examples)
- Fires on
- 0.61% of tokens
- Mostly in
- agent
- Layer
- after block 4
Where it fires
- agent 94%
- story 3%
- chat 2%
- python 1%
- worksheet 0%
Makes the model say
False us st meitictcontains
Makes it not say
+ equals *G told ==Fixed sum
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
agent 5.27, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a
-
agent 5.25, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a
-
agent 5.24, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a
-
agent 5.23, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a
-
agent 5.2, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a
-
agent 5.2, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a
-
agent 5.19, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a
-
agent 5.18, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a
-
agent 5.18, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a
-
agent 5.18, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a
-
agent 5.18, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a
-
agent 5.18, write, execute<|system_end|><|system_start|>You are an AI agent running in OpenCode, a