112 lines
2.8 KiB
Markdown
112 lines
2.8 KiB
Markdown
# HMMT26
|
||
|
||
## 概述
|
||
|
||
HMMT February 2026 是一个具有挑战性的评测基准,源自 2026 年 2 月举办的哈佛-麻省理工数学竞赛(Harvard-MIT Mathematics Tournament, HMMT),该赛事是全球最具声望且难度最高的高中数学竞赛之一。
|
||
|
||
## 任务描述
|
||
|
||
- **任务类型**:竞赛数学问题求解
|
||
- **输入**:HMMT 级别的数学问题
|
||
- **输出**:包含逐步推理的答案
|
||
- **难度**:高级高中竞赛水平
|
||
|
||
## 主要特点
|
||
|
||
- 包含 33 道来自 HMMT February 2026 竞赛的题目
|
||
- 覆盖四大核心领域:代数(Algebra)、组合数学(Combinatorics)、几何(Geometry)、数论(Number Theory)
|
||
- 题目难度极高,属于竞赛级别
|
||
- 考察高级数学推理能力
|
||
- 代表顶尖高中数学竞赛难度
|
||
|
||
## 评测说明
|
||
|
||
- 默认配置采用 **0-shot** 评测方式
|
||
- 答案需使用 `\boxed{}` 格式包裹,以便正确提取
|
||
- 使用数值精度指标,并结合符号等价性检查
|
||
- 题目涵盖多个数学领域
|
||
|
||
## 属性
|
||
|
||
| 属性 | 值 |
|
||
|----------|-------|
|
||
| **基准测试名称** | `hmmt26` |
|
||
| **数据集ID** | [evalscope/hmmt_feb_2026](https://modelscope.cn/datasets/evalscope/hmmt_feb_2026/summary) |
|
||
| **论文** | N/A |
|
||
| **标签** | `Math`, `Reasoning` |
|
||
| **指标** | `acc` |
|
||
| **默认示例数量** | 0-shot |
|
||
| **评测划分** | `train` |
|
||
|
||
## 数据统计
|
||
|
||
| 指标 | 值 |
|
||
|--------|-------|
|
||
| 总样本数 | 33 |
|
||
| 提示词长度(平均) | 385.3 字符 |
|
||
| 提示词长度(最小/最大) | 202 / 605 字符 |
|
||
|
||
## 样例示例
|
||
|
||
**子集**: `default`
|
||
|
||
```json
|
||
{
|
||
"input": [
|
||
{
|
||
"id": "02be53e7",
|
||
"content": "Problem:\nA line intersects the graph of \\( y = x^2 + \\frac{2}{x} \\) at three distinct points. Given that the \\( x \\)-coordinates of two of the points are 6 and 7, respectively, compute the \\( x \\)-coordinate of the third point.\n\nPlease reason step by step, and put your final answer within \\boxed{}.\n"
|
||
}
|
||
],
|
||
"target": "-\\frac{1}{21}",
|
||
"id": 0,
|
||
"group_id": 0,
|
||
"metadata": {
|
||
"problem_idx": 1,
|
||
"problem_type": [
|
||
"Algebra"
|
||
]
|
||
}
|
||
}
|
||
```
|
||
|
||
## 提示模板
|
||
|
||
**提示模板:**
|
||
```text
|
||
Problem:
|
||
{question}
|
||
|
||
Please reason step by step, and put your final answer within \boxed{{}}.
|
||
|
||
```
|
||
|
||
## 使用方法
|
||
|
||
### 使用 CLI
|
||
|
||
```bash
|
||
evalscope eval \
|
||
--model YOUR_MODEL \
|
||
--api-url OPENAI_API_COMPAT_URL \
|
||
--api-key EMPTY_TOKEN \
|
||
--datasets hmmt26 \
|
||
--limit 10 # 正式评测时请删除此行
|
||
```
|
||
|
||
### 使用 Python
|
||
|
||
```python
|
||
from evalscope import run_task
|
||
from evalscope.config import TaskConfig
|
||
|
||
task_cfg = TaskConfig(
|
||
model='YOUR_MODEL',
|
||
api_url='OPENAI_API_COMPAT_URL',
|
||
api_key='EMPTY_TOKEN',
|
||
datasets=['hmmt26'],
|
||
limit=10, # 正式评测时请删除此行
|
||
)
|
||
|
||
run_task(task_cfg=task_cfg)
|
||
``` |