Feature #4083
Fires on “⏎” (not labeled yet: read the examples)
- Fires on
- 0.37% of tokens
- Mostly in
- python
- Layer
- after block 4
Where it fires
- python 81%
- agent 11%
- story 6%
- chat 1%
- worksheet 1%
Makes the model say
unexpected trained$ test#P mini sorry
Makes it not say
import himself⏎ stace herselful
Strongest examples
On text the dictionary never trained on. The underlined token is where it fires most.
-
python 4.05def min_item(nums):⏎ return min(nums)⏎⏎$ python3 main
-
python 4.04# config.json⏎{"debug": false}⏎⏎# test_number
-
python 3.99"https://opencode.ai/config.json"⏎}⏎⏎# main.py
-
python 3.98⏎⏎⏎def quiet(s):⏎ return s.lower()⏎⏎# stats.py
-
python 3.95, 3]))⏎print(count_vowels("lemon"))⏎⏎$ python3 hello
-
python 3.94⏎def cubed(x):⏎ return x * x * x⏎⏎# main.py
-
python 3.9"aeiou":⏎ k = k + 1⏎ return k⏎⏎$ python3 demo
-
python 3.9nums):⏎ return sum(nums) / len(nums)⏎⏎$ python3 run
-
python 3.87()⏎⏎⏎def exclaim(s):⏎ return s + "!"⏎⏎# test_words
-
python 3.86py⏎def tail(items):⏎ return items[-1]⏎⏎# test_arith
-
python 3.85⏎ return len(s)⏎⏎# LICENSE⏎MIT License⏎⏎$ python3 hello
-
python 3.84# config.json⏎{"debug": false}⏎⏎$ python3 script