Update README.md

README.md

@@ -24,8 +24,22 @@ tags:
 
 **ECCV 2026**
 
+• [Project Page](https://janeyeon.github.io/hyflclip/)  
+• [Paper](https://arxiv.org/abs/2607.00428)  
+• [Code](https://github.com/janeyeon/hyfl-clip)
+
 </div>
 
+## Overview
+
+HyFL-CLIP improves the **robustness of long-context vision-language understanding** by transferring OpenCLIP’s pretrained Euclidean alignment into a **hyperbolic representation space**.
+The model reports improvements of up to **19.5%** in long-text cross-modal retrieval under perturbations including:
+
+- caption reordering;
+- caption summarization;
+- caption condensation; and
+- partial text deletion.
+
 ## Model Description
 
 **HyFL-CLIP** is a hyperbolic fine-tuning framework that improves the robustness of CLIP to long and compositional text descriptions.
@@ -52,24 +66,9 @@ Summarized tokens, short captions, long-form descriptions, and images are embedd
 
 Token- and phrase-level representations are aggregated using the Lorentzian centroid, also known as the Einstein midpoint. This provides a geometry-consistent method for combining partial representations into whole-level representations.
 
-The main hyperbolic operations include:
-
-- exponential and logarithmic maps at the origin;
-- Lorentzian inner products and distances;
-- parallel transport; and
-- Einstein midpoint aggregation.
 
 ## Results
 
-HyFL-CLIP improves long-text cross-modal retrieval and demonstrates greater robustness to textual perturbations.
-
-The model reports improvements of up to **19.5%** in long-text cross-modal retrieval under perturbations including:
-
-- caption reordering;
-- caption summarization;
-- caption condensation; and
-- partial text deletion.
-
 The evaluation covers:
 
 - long-context cross-modal retrieval;
@@ -312,7 +311,3 @@ Please cite the following paper when using HyFL-CLIP:
 }
 ```
 
-## Links
-
-- [Project Page](https://janeyeon.github.io/hyflclip/)
-- [Paper](https://arxiv.org/abs/2607.00428)