codybum commited on
Commit
1f84a3a
·
verified ·
1 Parent(s): 601d866

Commercial IFEval AVG4 baselines: tiered Verified/Reported/Estimate list with citations

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -49,8 +49,8 @@ less data**, two orders of magnitude smaller. It reaches the original-ChatGPT ba
49
  data and compute.
50
 
51
  **And it reaches that bar.** On **IFEval**, the standard instruction-following benchmark, KOS-V4-Instruct scores
52
- **AVG4 62.7**, above the verified **GPT-3.5-turbo (gpt-3.5-turbo-1106, 58.0)** and within about three points of
53
- a commercial 3B, **Qwen2.5-3B-Instruct (65.4)**.
54
 
55
  | IFEval (AVG4 of the four strict and loose metrics) | score | who built it, and how |
56
  | :-- | :-- | :-- |
 
49
  data and compute.
50
 
51
  **And it reaches that bar.** On **IFEval**, the standard instruction-following benchmark, KOS-V4-Instruct scores
52
+ **AVG4 62.7**, above the proprietary **ChatGPT (gpt-3.5-turbo-1106, 58.0)** and within three points of the
53
+ commercially trained 3B, **Qwen2.5-3B-Instruct (65.4)**.
54
 
55
  | IFEval (AVG4 of the four strict and loose metrics) | score | who built it, and how |
56
  | :-- | :-- | :-- |