diff --git a/.github/workflows/generate-html.yml b/.github/workflows/generate-html.yml
index 5f50de2..b2c8938 100644
--- a/.github/workflows/generate-html.yml
+++ b/.github/workflows/generate-html.yml
@@ -15,13 +15,12 @@ jobs:
     permissions:
       contents: write
       pull-requests: write
+      issues: write
     
     steps:
     - uses: actions/checkout@v3
       with:
-        ref: ${{ github.event.pull_request.head.ref }}
-        repository: ${{ github.event.pull_request.head.repo.full_name }}
-        token: ${{ secrets.GITHUB_TOKEN }}
+        ref: main
     
     - name: Set up Python
       uses: actions/setup-python@v4
@@ -37,18 +36,40 @@ jobs:
       run: |
         python src/generate.py awesome_3dgs_papers.yaml index.html
 
-    - name: Commit changes
-      if: github.event_name == 'pull_request'
+    - name: Create Update Branch and Commit Changes
       run: |
+        # Configure git
         git config --local user.email "41898282+github-actions[bot]@users.noreply.github.com"
         git config --local user.name "github-actions[bot]"
+        
+        # Create or update the branch
+        git checkout -B update-html
+        
+        # Stage and commit changes
         git add index.html
-        git diff --staged --quiet || (git commit -m "Auto-generate index.html" && git push)
+        if git diff --staged --quiet; then
+          echo "No changes to commit"
+          exit 0
+        fi
+        
+        # Commit and force push to update-html branch
+        git commit -m "Auto-generate index.html"
+        git push -f origin update-html
 
-    - name: Direct Push to Main
-      if: github.event_name == 'push'
+    - name: Create Pull Request
+      env:
+        GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
       run: |
-        git config --local user.email "41898282+github-actions[bot]@users.noreply.github.com"
-        git config --local user.name "github-actions[bot]"
-        git add index.html
-        git diff --staged --quiet || (git commit -m "Auto-generate index.html" && git push)
+        # Check if PR already exists
+        pr_exists=$(gh pr list --head update-html --base main --json number --jq 'length')
+        
+        if [ "$pr_exists" = "0" ]; then
+          # Create PR if it doesn't exist
+          gh pr create \
+            --title "Update generated HTML" \
+            --body "This is an automatically maintained PR that keeps the generated HTML in sync with YAML changes." \
+            --base main \
+            --head update-html || echo "Failed to create PR, may already exist"
+        else
+          echo "PR already exists, skipping creation"
+        fi
diff --git a/assets/thumbnails/liang2024wonderland.jpg b/assets/thumbnails/liang2024wonderland.jpg
new file mode 100644
index 0000000..811ae49
Binary files /dev/null and b/assets/thumbnails/liang2024wonderland.jpg differ
diff --git a/awesome_3dgs_papers.yaml b/awesome_3dgs_papers.yaml
index 1f3d623..55bc1f4 100644
--- a/awesome_3dgs_papers.yaml
+++ b/awesome_3dgs_papers.yaml
@@ -859,6 +859,42 @@
   - Project
   thumbnail: assets/thumbnails/taubner2024cap4d.jpg
   publication_date: '2024-12-16T18:58:51+00:00'
+- id: liang2024wonderland
+  title: 'Wonderland: Navigating 3D Scenes from a Single Image'
+  authors: Hanwen Liang, Junli Cao, Vidit Goel, Guocheng Qian, Sergei Korolev, Demetri
+    Terzopoulos, Konstantinos N. Plataniotis, Sergey Tulyakov, Jian Ren
+  year: '2024'
+  abstract: 'This paper addresses a challenging question: How can we efficiently create
+    high-quality, wide-scope 3D scenes from a single arbitrary image? Existing methods
+    face several constraints, such as requiring multi-view data, time-consuming per-scene
+    optimization, low visual quality in backgrounds, and distorted reconstructions
+    in unseen areas. We propose a novel pipeline to overcome these limitations. Specifically,
+    we introduce a large-scale reconstruction model that uses latents from a video
+    diffusion model to predict 3D Gaussian Splattings for the scenes in a feed-forward
+    manner. The video diffusion model is designed to create videos precisely following
+    specified camera trajectories, allowing it to generate compressed video latents
+    that contain multi-view information while maintaining 3D consistency. We train
+    the 3D reconstruction model to operate on the video latent space with a progressive
+    training strategy, enabling the efficient generation of high-quality, wide-scope,
+    and generic 3D scenes. Extensive evaluations across various datasets demonstrate
+    that our model significantly outperforms existing methods for single-view 3D scene
+    generation, particularly with out-of-domain images. For the first time, we demonstrate
+    that a 3D reconstruction model can be effectively built upon the latent space
+    of a diffusion model to realize efficient 3D scene generation.
+
+    '
+  project_page: https://snap-research.github.io/wonderland/
+  paper: https://arxiv.org/pdf/2412.12091v1.pdf
+  code: null
+  video: null
+  tags:
+  - Feed-Forward
+  - Project
+  - Sparse
+  - World Generation
+  thumbnail: assets/thumbnails/liang2024wonderland.jpg
+  publication_date: '2024-12-16T18:58:17+00:00'
+  date_source: arxiv
 - id: tang2024gaf
   title: 'GAF: Gaussian Avatar Reconstruction from Monocular Videos via Multi-view
     Diffusion'