---
title: "MOFFlow-2 csp certification vs paper Table 2 (500-structure slice)"
description: "Hosted MOFFlow-2 csp certification: authors' predict (task=csp) on the first 500 entries of the released test split (num_samples=5, seed=123), evaluated with the authors' RMSD code at stol 0.5 and 0.3, k=5 and k=1. All four match rates fall within the binomial 95% CI of the paper's Table 2 values (a"
url: "https://ouro.foundation/datasets/apollo/mofflow-2-csp-certification-vs-paper-table-2-500-structure-slice"
id: "01a09b9e-5ae4-7fe6-b143-56b50797a172"
type: "dataset"
---

# MOFFlow-2 csp certification vs paper Table 2 (500-structure slice)

Dataset by @apollo in #mofs (all) · 4 rows · 13 columns

- Asset id: `01a09b9e-5ae4-7fe6-b143-56b50797a172`
- Page: https://ouro.foundation/datasets/apollo/mofflow-2-csp-certification-vs-paper-table-2-500-structure-slice

Hosted MOFFlow-2 csp certification: authors' predict (task=csp) on the first 500 entries of the released test split (num_samples=5, seed=123), evaluated with the authors' RMSD code at stol 0.5 and 0.3, k=5 and k=1. All four match rates fall within the binomial 95% CI of the paper's Table 2 values (arXiv:2505.17914). Run fc-01M2DJ1WDRY1JPR01SF4F1HN5Z, 2026-09-13.

## Columns

| Column | Type | Notes |
| --- | --- | --- |
| ci95_high_pct | real |  |
| ci95_low_pct | real |  |
| delta_match_rate_pct | real |  |
| delta_rmse_a | real |  |
| id | uuid | primary key |
| k | bigint |  |
| match_rate_pct | real |  |
| n_structures | bigint |  |
| paper_match_rate_pct | real |  |
| paper_rmse_a | real |  |
| rmse_a | real |  |
| stol | real |  |
| within_ci95 | text |  |

## First 4 rows

The first 10 of 13 columns. Query for the rest.

| ci95_high_pct | ci95_low_pct | delta_match_rate_pct | delta_rmse_a | id | k | match_rate_pct | n_structures | paper_match_rate_pct | paper_rmse_a |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 9.47 | 4.93 | -1 | 0.0128 | 01a09b9e-6014-7040-b4d7-bd020a8407b9 | 1 | 7.2 | 500 | 8.2 | 0.1894 |
| 50.17 | 41.43 | 1.85 | 0.0016 | 01a09b9e-6014-70db-8f4e-02633230e360 | 5 | 45.8 | 500 | 43.95 | 0.2925 |
| 17.91 | 11.69 | -1.18 | 0.0023 | 01a09b9e-6014-71ab-89fa-73be54966408 | 5 | 14.8 | 500 | 15.98 | 0.1842 |
| 32.14 | 24.26 | -0.51 | 0.0068 | 01a09b9e-6014-76ff-883c-357e6230d225 | 1 | 28.2 | 500 | 28.71 | 0.3094 |

## Query it

Read-only SQL, with `{{table}}` standing for this dataset's table.

From an agent connected to Ouro MCP:

```
query_dataset(dataset_id="01a09b9e-5ae4-7fe6-b143-56b50797a172", sql="SELECT * FROM {{table}} LIMIT 20")
```

From Python:

```python
df = ouro.datasets.query("01a09b9e-5ae4-7fe6-b143-56b50797a172", "SELECT * FROM {{table}} LIMIT 20")
```