Upload folder using huggingface_hub

README.md

@@ -16,6 +16,14 @@ size_categories:
 
 # EvasionBench
 
+<p align="center">
+  <a href="https://iiiiqiiii.github.io/EvasionBench"><img src="https://img.shields.io/badge/Project-Page-blue?style=for-the-badge" alt="Project Page"></a>
+  <a href="https://huggingface.co/FutureMa/Eva-4B-V2"><img src="https://img.shields.io/badge/🤗-Model-yellow?style=for-the-badge" alt="Model"></a>
+  <a href="https://github.com/IIIIQIIII/EvasionBench"><img src="https://img.shields.io/badge/GitHub-Repo-black?style=for-the-badge&logo=github" alt="GitHub"></a>
+  <a href="https://colab.research.google.com/github/IIIIQIIII/EvasionBench/blob/main/scripts/eva4b_inference.ipynb"><img src="https://img.shields.io/badge/Colab-Quick_Start-F9AB00?style=for-the-badge&logo=googlecolab" alt="Open In Colab"></a>
+  <a href="https://arxiv.org/abs/2602.xxxxx"><img src="https://img.shields.io/badge/arXiv-Paper-red?style=for-the-badge" alt="Paper"></a>
+</p>
+
 EvasionBench is a benchmark dataset for detecting **evasive answers** in earnings call Q&A sessions. The task is to classify how directly corporate management addresses questions from financial analysts.
 
 ## Dataset Description
@@ -26,7 +34,7 @@ EvasionBench is a benchmark dataset for detecting **evasive answers** in earning
 
 ### Dataset Summary
 
-This dataset contains 16,726 question-answer pairs from earnings call transcripts, each labeled with one of three evasion levels. The labels were generated using the Eva-4B model, a fine-tuned classifier specifically trained for financial discourse evasion detection.
+This dataset contains 16,726 question-answer pairs from earnings call transcripts, each labeled with one of three evasion levels. The labels were generated using the Eva-4B-V2 model, a fine-tuned classifier specifically trained for financial discourse evasion detection.
 
 ### Supported Tasks
 
@@ -82,7 +90,7 @@ English
 ```python
 from datasets import load_dataset
 
-dataset = load_dataset("path/to/EvasionBench")
+dataset = load_dataset("FutureMa/EvasionBench")
 ```
 
 ### Loading from Parquet
@@ -93,6 +101,43 @@ import pandas as pd
 df = pd.read_parquet("evasionbench_17k_eva4b_labels_dedup.parquet")
 ```
 
+### Quick Start: Inference with Eva-4B-V2
+
+````python
+from datasets import load_dataset
+from transformers import pipeline
+
+# Load dataset
+dataset = load_dataset("FutureMa/EvasionBench")
+
+# Load model using text-generation pipeline
+pipe = pipeline("text-generation", model="FutureMa/Eva-4B-V2", device_map="auto")
+
+# Get a sample
+sample = dataset["train"][0]
+question = sample["question"]
+answer = sample["answer"]
+
+# Prepare prompt
+prompt = f"""You are a financial analyst. Your task is to Detect Evasive Answers in Financial Q&A
+
+Question: {question}
+Answer: {answer}
+
+Response format:
+```json
+{{"label": "direct|intermediate|fully_evasive"}}
+```
+
+Answer in ```json content, no other text"""
+
+# Run inference
+result = pipe(prompt, max_new_tokens=64, do_sample=False)
+print(result[0]["generated_text"])
+````
+
+For a complete inference example with batch processing and evaluation, see our [Colab notebook](https://colab.research.google.com/github/IIIIQIIII/EvasionBench/blob/main/scripts/eva4b_inference.ipynb).
+
 ## Dataset Creation
 
 ### Source Data
@@ -101,7 +146,7 @@ The question-answer pairs are derived from publicly available earnings call tran
 
 ### Annotation Process
 
-Labels were generated using **Eva-4B**, a 4B parameter model fine-tuned specifically for evasion detection in financial discourse. Eva-4B achieves 84.9% Macro-F1 on the EvasionBench evaluation set, outperforming frontier LLMs including Claude Opus 4.5 and Gemini 3 Flash.
+Labels were generated using **Eva-4B-V2**, a 4B parameter model fine-tuned specifically for evasion detection in financial discourse. Eva-4B-V2 achieves 84.9% Macro-F1 on the EvasionBench evaluation set, outperforming frontier LLMs including Claude Opus 4.5 and Gemini 3 Flash.
 
 ## Considerations for Using the Data
 
@@ -114,7 +159,7 @@ This dataset can be used to:
 
 ### Limitations
 
-- Labels are model-generated (Eva-4B) rather than human-annotated
+- Labels are model-generated (Eva-4B-V2) rather than human-annotated
 - The dataset reflects the distribution of evasion patterns in the source data
 - Performance may vary across different industries or time periods