| | --- |
| | license: apache-2.0 |
| | language: |
| | - en |
| | size_categories: |
| | - n<1K |
| | task_categories: |
| | - question-answering |
| | - multiple-choice |
| | configs: |
| | - config_name: benchmark |
| | data_files: |
| | - split: test |
| | path: dataset.json |
| | tags: |
| | - geospatial |
| | annotations_creators: |
| | - expert-generated |
| | paperswithcode_id: mapeval-textual |
| | --- |
| | |
| |
|
| | # MapEval-Textual |
| |
|
| | [MapEval](https://arxiv.org/abs/2501.00316)-Textual is created using [MapQaTor](https://arxiv.org/abs/2412.21015). |
| |
|
| | ## Usage |
| |
|
| | ```python |
| | from datasets import load_dataset |
| | |
| | # Load dataset |
| | ds = load_dataset("MapEval/MapEval-Textual", name="benchmark") |
| | |
| | # Generate better prompts |
| | for item in ds["test"]: |
| | # Start with a clear task description |
| | prompt = ( |
| | "You are a highly intelligent assistant. " |
| | "Based on the given context, answer the multiple-choice question by selecting the correct option.\n\n" |
| | "Context:\n" + item["context"] + "\n\n" |
| | "Question:\n" + item["question"] + "\n\n" |
| | "Options:\n" |
| | ) |
| | |
| | # List the options more clearly |
| | for i, option in enumerate(item["options"], start=1): |
| | prompt += f"{i}. {option}\n" |
| | |
| | # Add a concluding sentence to encourage selection of the answer |
| | prompt += "\nSelect the best option by choosing its number." |
| | |
| | # Use the prompt as needed |
| | print(prompt) # Replace with your processing logic |
| | ``` |
| |
|
| | ## Leaderboard |
| |
|
| | | Model | Overall | Place Info | Nearby | Routing | Trip | Unanswerable | |
| | |--------------------------|:---------:|:------------:|:--------:|:---------:|:--------:|:--------------:| |
| | | Claude-3.5-Sonnet | **66.33** | **73.44** | 73.49 | **75.76** | **49.25** | 40.00 | |
| | | Gemini-1.5-Pro | **66.33** | 65.63 | **74.70** | 69.70 | 47.76 | **85.00** | |
| | | GPT-4o | 63.33 | 64.06 | **74.70** | 69.70 | **49.25** | 40.00 | |
| | | GPT-4-Turbo | 62.33 | 67.19 | 71.08 | 71.21 | 47.76 | 30.00 | |
| | | Gemini-1.5-Flash | 58.67 | 62.50 | 67.47 | 66.67 | 38.81 | 50.00 | |
| | | GPT-4o-mini | 51.00 | 46.88 | 63.86 | 57.58 | 40.30 | 25.00 | |
| | | GPT-3.5-Turbo | 37.67 | 26.56 | 53.01 | 48.48 | 28.36 | 5.00 | |
| | | Llama-3.1-70B | 61.00 | 70.31 | 67.47 | 69.70 | 40.30 | 45.00 | |
| | | Llama-3.2-90B | 58.33 | 68.75 | 66.27 | 66.67 | 38.81 | 30.00 | |
| | | Qwen2.5-72B | 57.00 | 62.50 | 71.08 | 63.64 | 41.79 | 10.00 | |
| | | Qwen2.5-14B | 53.67 | 57.81 | 71.08 | 59.09 | 32.84 | 20.00 | |
| | | Gemma-2.0-27B | 49.00 | 39.06 | 71.08 | 59.09 | 31.34 | 15.00 | |
| | | Gemma-2.0-9B | 47.33 | 50.00 | 50.60 | 59.09 | 34.33 | 30.00 | |
| | | Llama-3.1-8B | 44.00 | 53.13 | 57.83 | 45.45 | 23.88 | 20.00 | |
| | | Qwen2.5-7B | 43.33 | 48.44 | 49.40 | 42.42 | 38.81 | 20.00 | |
| | | Mistral-Nemo | 43.33 | 46.88 | 50.60 | 50.00 | 32.84 | 15.00 | |
| | | Mixtral-8x7B | 43.00 | 53.13 | 54.22 | 45.45 | 26.87 | 10.00 | |
| | | Phi-3.5-mini | 37.00 | 40.63 | 48.19 | 46.97 | 20.90 | 0.00 | |
| | | Llama-3.2-3B | 33.00 | 31.25 | 49.40 | 31.82 | 25.37 | 0.00 | |
| | | Human | 86.67 | 92.19 | 90.36 | 81.81 | 88.06 | 65.00 | |
| |
|
| |
|
| | ## Citation |
| |
|
| | If you use this dataset, please cite the original paper: |
| |
|
| | ``` |
| | @article{dihan2024mapeval, |
| | title={MapEval: A Map-Based Evaluation of Geo-Spatial Reasoning in Foundation Models}, |
| | author={Dihan, Mahir Labib and Hassan, Md Tanvir and Parvez, Md Tanvir and Hasan, Md Hasebul and Alam, Md Almash and Cheema, Muhammad Aamir and Ali, Mohammed Eunus and Parvez, Md Rizwan}, |
| | journal={arXiv preprint arXiv:2501.00316}, |
| | year={2024} |
| | } |
| | ``` |
