You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 4e2fdcc
Browse filesBrowse the repository at this point in the historyBrowse files
Copy file name to clipboardExpand all lines: README.md
+2-2Lines changed: 2 additions & 2 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -26,9 +26,9 @@ Get involved! Join our [discord server](https://discord.gg/KJySXsrSX5).
26
26
27
27
## Introduction
28
28
29
-
Coding capability in LLMs is trained using techniques like [Inverse-Instruct](https://www.researchgate.net/figure/The-overview-of-INVERSE-INSTRUCT-INVERSE-INSTRUCT-utilizes-the-models-own-capability-in_fig1_382079949)where exisiting model acts as a summarizer, looking at the code and writing instructions that could realistically output that exact code block.
29
+
Coding capability in LLMs is trained using techniques like [Inverse-Instruct](https://www.researchgate.net/figure/The-overview-of-INVERSE-INSTRUCT-INVERSE-INSTRUCT-utilizes-the-models-own-capability-in_fig1_382079949)with exisiting model acting as a summarizer, looking at the code and writing instructions that could realistically output that exact code block.
30
30
31
-
CWC was designed to tap directly into these summarizations by showing the model a complete set of files, together with your instructions from the start.
31
+
By showing the model a complete set of files, together with your instructions from the start, CWC was designed to tap directly into these snippet-summarization pairs.
32
32
33
33
This approach encourages intention in file selection, which makes simple instructions just work and the required model's reasoning much shorter or even optional.
0 commit comments