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
- Changed 'we/our' to 'I/my' throughout documentation
- Properly credit Steve Solun as sole author in citation
- Updated GitHub URL to stevesolun/Chameleon
- Removed team references - this is individual work
- Fixed 'research collaborators' to 'community feedback'
- Ensured all language reflects single-author contribution
Copy file name to clipboardExpand all lines: README.md
+6-6Lines changed: 6 additions & 6 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -8,7 +8,7 @@ A comprehensive framework for testing large language model robustness under lexi
8
8
9
9
## 🎯 Project Overview
10
10
11
-
The Chameleon framework addresses a critical question in AI evaluation: **How robust are large language models to lexical variations in input text?** By applying controlled lexical distortions to academic assessment questions, we can measure model performance degradation and identify vulnerable domains.
11
+
The Chameleon framework addresses a critical question in AI evaluation: **How robust are large language models to lexical variations in input text?** By applying controlled lexical distortions to academic assessment questions, I can measure model performance degradation and identify vulnerable domains.
Our comprehensive statistical analysis using **McNemar's test for paired comparisons** reveals statistically significant performance degradation across multiple dimensions:
378
+
My comprehensive statistical analysis using **McNemar's test for paired comparisons** reveals statistically significant performance degradation across multiple dimensions:
379
379
380
380
#### 🎯 **Key Statistical Findings**
381
381
-**100% of distortion levels** show highly significant degradation (p < 0.001)
0 commit comments