Feature #3283
Fires on “ you” (not labeled yet: read the examples)
- Fires on
- 0.38% of tokens
- Mostly in
- chat
- Layer
- after block 4
Where it fires
- chat 42%
- story 26%
- agent 19%
- worksheet 9%
- python 3%
Makes the model say
path tail=**/script were defines generator whether
Makes it not say
⏎. H(" you ini went
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
chat 5.38and fix the Python files in this project.<|assistant_end|><|bos|><|user_start|>What can you do?<|user_end|><|assistant_start|>
-
chat 5.14and fix the Python files in this project.<|assistant_end|><|bos|><|user_start|>What can you do?<|user_end|><|assistant_start|>
-
chat 5.13tell short stories and add numbers.<|assistant_end|><|user_start|>What can you do?<|user_end|><|assistant_start|>
-
chat 5.1288.<|assistant_end|><|user_start|>Can you add 55
-
chat 5.04a helpful assistant.<|system_end|><|user_start|>Can you add 53
-
chat 4.88story!<|assistant_end|><|user_start|>Can you book a table for
-
chat 4.83<|user_start|>Can you tell me a story
-
agent 4.79agent running in OpenCode, a coding agent harness.<|system_end|><|user_start|>"Can you run the tests?"
-
agent 4.72agent running in OpenCode, a coding agent harness.<|system_end|><|user_start|>"Can you run main.py
-
agent 4.69agent running in OpenCode, a coding agent harness.<|system_end|><|user_start|>"Can you run demo.py
-
agent 4.68agent running in OpenCode, a coding agent harness.<|system_end|><|user_start|>"Can you run hello.py
-
agent 4.68agent running in OpenCode, a coding agent harness.<|system_end|><|user_start|>"Can you run the tests?"