Feature #1296
Fires on “ an” (not labeled yet: read the examples)
- Fires on
- 0.36% of tokens
- Mostly in
- agent
- Layer
- after block 4
Where it fires
- agent 93%
- python 3%
- chat 2%
- story 2%
- worksheet 0%
Makes the model say
exclamationeven right) runs please)))`
Makes it not say
⏎8If not They5illnd
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
agent 3.07cute, context7_query<|system_end|><|system_start|>You are an AI agent running in
-
agent 3.06search, write, execute<|system_end|><|system_start|>You are an AI agent running in
-
agent 3.06search, write, execute<|system_end|><|system_start|>You are an AI agent running in
-
agent 3.03search, write, execute<|system_end|><|system_start|>You are an AI agent running in
-
agent 3.03search, write, execute<|system_end|><|system_start|>You are an AI agent running in
-
agent 3.01search, write, execute<|system_end|><|system_start|>You are an AI agent running in
-
agent 3.01search, write, execute<|system_end|><|system_start|>You are an AI agent running in
-
agent 3.0search, write, execute<|system_end|><|system_start|>You are an AI agent running in
-
agent 2.99search, write, execute<|system_end|><|system_start|>You are an AI agent running in
-
agent 2.99cute, context7_query<|system_end|><|system_start|>You are an AI agent running in
-
agent 2.99search, write, execute<|system_end|><|system_start|>You are an AI agent running in
-
agent 2.98search, write, execute<|system_end|><|system_start|>You are an