File size: 6,848 Bytes
457e886
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d4c3c2d
457e886
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
---
language:
- en
license:
- cc-by-4.0
- cc-by-nc-sa-4.0
task_categories:
- visual-question-answering
- multiple-choice
pretty_name: RFSchemBench
size_categories:
- 1K<n<10K
configs:
- config_name: permissive
  default: true
  data_files:
  - split: test
    path: data/permissive/test-*.parquet
- config_name: nc_allowed
  data_files:
  - split: test
    path: data/nc_allowed/test-*.parquet
tags:
- rf
- circuit
- schematic
- multimodal
- electronic-engineering
- benchmark
- vqa
---

# RFSchemBench

A multimodal LLM evaluation benchmark for **radio-frequency circuit schematic understanding**, organized by a four-level semantic hierarchy:

1. **Component Understanding** — visible component, parameter, label, and supply-rail recognition.
2. **Structural Understanding** — net membership, pin-to-net mapping, boundary connectivity, and pair-via-net topological reasoning.
3. **Functional Understanding** — circuit functional role, signal-form classification, supply strategy, sub-type identification.
4. **Dynamic Reasoning** — counterfactual plot choice and schematic-modification ↔ simulation-result matching, grounded in `ngspice` simulation.

The benchmark contains **2,348 questions across 590 rendered schematic pages** from publicly available RF schematic data.

## Quick start

```python
from datasets import load_dataset

# Permissive subset (CC-BY-4.0; recommended for most users)
ds = load_dataset("anonymous-submission042/RFSchemBench", "permissive", split="test")

# Full benchmark including a NonCommercial-ShareAlike subset
ds_nc = load_dataset("anonymous-submission042/RFSchemBench", "nc_allowed", split="test")

print(ds[0]["question"], "→ answer:", ds[0]["answer"])
ds[0]["image"].show()  # PIL.Image of the schematic
```

## Configurations

| Config | Rows | License | Notes |
|---|---:|---|---|
| `permissive` (default) | 2,258 | `CC-BY-4.0` | Excludes the NC-licensed source class. Suitable for commercial / industrial reviewers. |
| `nc_allowed` | 2,348 | mixed `CC-BY-4.0` + `CC-BY-NC-SA-4.0` | Full benchmark. Per-row `license` field marks which items are NC-licensed. NonCommercial usage only. |

## Schema

Each row has the following fields:

| Field | Type | Description |
|---|---|---|
| `question_id` | string | Unique identifier (stable across releases) |
| `item_id` | string | Source schematic identifier |
| `source` | string | Source class (`qucs` / `kicad` / `myriadrf` / `m17` / `oresat`) |
| `level` | string | One of `Component Understanding` / `Structural Understanding` / `Functional Understanding` / `Dynamic Reasoning` |
| `category` | string | Coarse-grained tag |
| `question` | string | English prompt (what models are evaluated on) |
| `image` | PIL.Image | Primary schematic rendering (`image.png`) |
| `context_images` | list of `{caption, image}` | Auxiliary context images (Dynamic Reasoning only — schematic plus baseline / variant simulation plots) |
| `options` | list of `{label, text, image}` | Multi-choice options (Dynamic Reasoning only). Some options have only `text`, others have both `text` and `image`. |
| `answer_type` | string | `enum_label` / `comma_separated_list` / `integer` / `short_text` |
| `answer_allowed` | list of string | Permitted enum values (empty for non-enum types) |
| `answer` | string | Gold answer; for list-type answers, comma-separated |
| `source_schematic` | string | Provenance: original `.kicad_sch` / `.sch` path |
| `license` | string | Per-row license tag (`CC-BY-4.0` or `CC-BY-NC-SA-4.0`) |

## Construction

The benchmark is constructed via **expert-rule-guided programmatic generation from authoritative sources**:

- Domain experts encode question-generation rules and gold-answer semantics into Python programs.
- Gold answers are extracted deterministically from authoritative source artifacts (KiCad CLI output, Qucs native schematic graph, `ngspice` simulation outputs).
- LLMs are deliberately **excluded from the gold-answer path**; they are used only as an auxiliary RF-relevance gate at the page level.
- An iterative rule-refinement loop catches edge cases during construction; the released gold answers reflect the latest revisions.

This avoids the gold-answer noise floor of LLM-as-Generator benchmarks while scaling beyond purely human-curated efforts.

## License

This dataset is released under a **two-tier license model** because the upstream sources have heterogeneous licenses:

- **`permissive` config** (recommended default): all rows under `CC-BY-4.0`. Compatible with commercial use, redistribution, and derivative works subject to attribution.
- **`nc_allowed` config**: includes one source class (`m17` digital-radio community hardware, 90 questions) which is upstream-licensed under `CC-BY-NC-SA-4.0` (NonCommercial-ShareAlike). Per-row `license` field marks affected items. Users must respect NC + ShareAlike for those rows.

Per-source licensing summary:

| Source class | Upstream license profile | Tier inclusion |
|---|---|---|
| `qucs` | GPL-2.0 example schematics (treated as derivative-work CC-BY-4.0 for image renderings) | both |
| `kicad` | mostly MIT / Apache-2.0 / GPL-3.0 mix | both |
| `myriadrf` | mostly Apache-2.0 / CC-BY-4.0 | both |
| `oresat` | CERN-OHL-S-2.0 (treated as share-alike-compatible CC-BY-4.0 for renderings) | both |
| `m17` | **CC-BY-NC-SA-4.0** ⚠ NC | `nc_allowed` only |

For redistribution that requires fully permissive licensing, use only the `permissive` config.

## Limitations

1. **Source-class size imbalance**: question counts per source class span 40–974; per-source claims should be reported with N.
2. **Dynamic Reasoning scope**: only one source class has the simulation-grounded subset (55 questions). This dimension is reported as a small stress test, not the main result.
3. **Language**: questions are evaluated in English. (A Chinese parallel set was used internally during construction for human review but is not part of the released schema.)
4. **Single-image protocol**: each question is paired with one primary schematic image (Dynamic Reasoning rows additionally provide context plots / option plots).
5. **Anonymized release**: this submission account is for double-blind peer review. The dataset will be transferred to the official maintainer account upon acceptance.

## Citation

```bibtex
@misc{rfschembench2026,
  title  = {RFSchemBench: A Multi-Source, Hierarchically-Structured Multimodal Benchmark for RF Circuit Schematic Understanding},
  author = {Anonymous},
  year   = {2026},
  note   = {Submitted to NeurIPS 2026 Evaluations \& Datasets Track}
}
```

## Contact

For benchmark integrity issues (gold-answer corrections, RF-gate disputes, parser / scorer concerns), please open a Discussion on this dataset's HuggingFace page. During the double-blind review window, identifying contact details are intentionally withheld.