13274243a0
Keep K3 suite selection and report-schema scoring in bash, merge K3/vision dataset_args into dpv4 yamls, and pin EvalScope at 735d920ee911 with local patches. Co-authored-by: Cursor <cursoragent@cursor.com>
115 lines
3.0 KiB
Markdown
115 lines
3.0 KiB
Markdown
# HMMT-Nov-2025
|
|
|
|
|
|
## Overview
|
|
|
|
HMMT November 2025 (MathArena) is a challenging evaluation benchmark derived from the Harvard-MIT Mathematics Tournament (HMMT) November 2025 competition, one of the most prestigious and difficult high school math contests globally. It is a different contest from HMMT February 2025 (`hmmt25`).
|
|
|
|
## Task Description
|
|
|
|
- **Task Type**: Competition Mathematics Problem Solving
|
|
- **Input**: HMMT-level mathematical problem
|
|
- **Output**: Answer with step-by-step reasoning
|
|
- **Domain**: Algebra, Combinatorics, Geometry, and Number Theory
|
|
|
|
## Key Features
|
|
|
|
- 30 problems from the HMMT November 2025 competition
|
|
- Sourced from the MathArena `hmmt_nov_2025` dataset and mirrored on ModelScope
|
|
- Highly challenging competition-level problems
|
|
- Tests advanced mathematical reasoning
|
|
- Represents elite high school mathematics difficulty
|
|
|
|
## Evaluation Notes
|
|
|
|
- Default configuration loads `evalscope/hmmt_nov_2025` from ModelScope and evaluates the `train` split
|
|
- Default configuration uses **0-shot** evaluation
|
|
- Answers should be formatted within `\boxed{}` for proper extraction
|
|
- Numeric accuracy uses mathematical equivalence checking for integers, fractions, decimals, and symbolic expressions
|
|
- No additional runtime dependencies are required
|
|
|
|
|
|
## Properties
|
|
|
|
| Property | Value |
|
|
|----------|-------|
|
|
| **Benchmark Name** | `hmmt_nov25` |
|
|
| **Dataset ID** | [evalscope/hmmt_nov_2025](https://modelscope.cn/datasets/evalscope/hmmt_nov_2025/summary) |
|
|
| **Paper** | N/A |
|
|
| **Tags** | `Math`, `Reasoning` |
|
|
| **Metrics** | `accuracy` |
|
|
| **Default Shots** | 0-shot |
|
|
| **Evaluation Split** | `train` |
|
|
|
|
|
|
## Data Statistics
|
|
|
|
| Metric | Value |
|
|
|--------|-------|
|
|
| Total Samples | 30 |
|
|
| Prompt Length (Mean) | 403.3 chars |
|
|
| Prompt Length (Min/Max) | 203 / 759 chars |
|
|
|
|
## Sample Example
|
|
|
|
**Subset**: `default`
|
|
|
|
```json
|
|
{
|
|
"input": [
|
|
{
|
|
"id": "e53f11c5",
|
|
"content": "Problem:\nLet $ABCD$ be a rectangle. Let $X$ and $Y$ be points on segments $\\overlien{BC}$ and $\\overline{AD}$, respectively, such that $\\angle AXY = \\angle XYC = 90^\\circ$. Given that $AX : XY : YC = 1 : 2 : 1$ and $AB = 1$, compute $BC$.\n\nPlease reason step by step, and put your final answer within \\boxed{}.\n"
|
|
}
|
|
],
|
|
"target": "3",
|
|
"id": 0,
|
|
"group_id": 0,
|
|
"metadata": {
|
|
"problem_idx": 1,
|
|
"problem_type": null
|
|
}
|
|
}
|
|
```
|
|
|
|
## Prompt Template
|
|
|
|
**Prompt Template:**
|
|
```text
|
|
Problem:
|
|
{question}
|
|
|
|
Please reason step by step, and put your final answer within \boxed{{}}.
|
|
|
|
```
|
|
|
|
## Usage
|
|
|
|
### Using CLI
|
|
|
|
```bash
|
|
evalscope eval \
|
|
--model YOUR_MODEL \
|
|
--api-url OPENAI_API_COMPAT_URL \
|
|
--api-key EMPTY_TOKEN \
|
|
--datasets hmmt_nov25 \
|
|
--limit 10 # Remove this line for formal evaluation
|
|
```
|
|
|
|
### Using 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=['hmmt_nov25'],
|
|
limit=10, # Remove this line for formal evaluation
|
|
)
|
|
|
|
run_task(task_cfg=task_cfg)
|
|
```
|