Datasets:
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -27,81 +27,50 @@ configs:
|
|
| 27 |
path: qrels.jsonl
|
| 28 |
---
|
| 29 |
|
| 30 |
-
# GitBench: Real-World GitHub Pull Request Benchmark for AI Coding Agents
|
| 31 |
|
| 32 |
-
**GitBench** (originally *GitState-Bench*) is an open evaluation benchmark developed by the **Knot AI** team ([@Knot-Repo](https://github.com/Knot-Repo)) to rigorously evaluate how AI coding assistants, code search systems, and Retrieval-Augmented Generation (RAG) engines handle **temporal code evolution, obsolete
|
| 33 |
|
| 34 |
---
|
| 35 |
|
| 36 |
## Benchmark Motivation
|
| 37 |
|
| 38 |
-
Most coding benchmarks (e.g. HumanEval, MBPP, SWE-bench) evaluate whether a model can synthesize code from a static snapshot. However, in production engineering environments, developer context
|
| 39 |
|
| 40 |
-
AI agents regularly fail in three
|
| 41 |
-
1. **Temporal Obsolescence (
|
| 42 |
-
2. **Brainstorm / Draft Contamination**
|
| 43 |
-
3. **
|
| 44 |
-
|
| 45 |
-
GitBench benchmarks these failure modes on real merged Pull Requests scraped from **25 top Python (pip) & AI engineering repositories**.
|
| 46 |
|
| 47 |
---
|
| 48 |
|
| 49 |
## Dataset Scale & Repository Coverage
|
| 50 |
|
| 51 |
-
- **Total Corpus Documents:** `1,788` (`447` merged active
|
| 52 |
-
- **Developer Queries:** `447` natural
|
| 53 |
- **Relevance Judgments (`qrels`):** `1,788` graded relevance pairs (`4,023` total records across splits)
|
| 54 |
- **25 Sampled Repositories:**
|
| 55 |
`fastapi/fastapi`, `pydantic/pydantic`, `anthropics/anthropic-sdk-python`, `vercel/ai`, `encode/httpx`, `psf/requests`, `pallets/flask`, `tiangolo/sqlmodel`, `openai/openai-python`, `langchain-ai/langchain`, `run-llama/llama_index`, `astral-sh/uv`, `huggingface/transformers`, `vllm-project/vllm`, `pandas-dev/pandas`, `numpy/numpy`, `scikit-learn/scikit-learn`, `pytorch/pytorch`, `sqlalchemy/sqlalchemy`, `celery/celery`, `pytest-dev/pytest`, `Textualize/rich`, `streamlit/streamlit`, `gradio-app/gradio`, `BerriAI/litellm`.
|
| 56 |
|
| 57 |
---
|
| 58 |
|
| 59 |
-
##
|
| 60 |
-
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
- **`id`**: Unique document identifier (e.g., `fastapi_fastapi_pr_11520_active_gold`).
|
| 66 |
-
- **`repo`**: Source GitHub repository (e.g., `fastapi/fastapi`, `pydantic/pydantic`).
|
| 67 |
-
- **`pr_number`**: GitHub Pull Request number.
|
| 68 |
-
- **`pr_url`**: Verifiable public link to the original GitHub PR.
|
| 69 |
-
- **`doc_type`**: One of four semantic classes:
|
| 70 |
-
- `active_v2_gold`: Real merged PR implementation (the ground-truth current behavior).
|
| 71 |
-
- `superseded_v1`: Pre-PR legacy implementation superseded by this PR.
|
| 72 |
-
- `brainstorm_unconfirmed`: Unmerged / draft proposal on the same repository.
|
| 73 |
-
- `admin_secret_trap`: Synthetic CI/CD deploy token config designed to test leakage prevention.
|
| 74 |
-
- **`title`**: Descriptive title.
|
| 75 |
-
- **`text`**: Full context snippet.
|
| 76 |
-
|
| 77 |
-
### 2. `queries`
|
| 78 |
-
Natural developer questions targeting specific features:
|
| 79 |
-
- **`query_id`**: Unique query identifier.
|
| 80 |
-
- **`repo`**: Repository under question.
|
| 81 |
-
- **`pr_number`**: Targeted PR.
|
| 82 |
-
- **`text`**: Developer question (e.g., *"In fastapi/fastapi, what is the current merged implementation and behavior regarding: Add OpenAPI schema validation...?"*).
|
| 83 |
-
- **`gold_id`**: Ground truth active document ID.
|
| 84 |
-
- **`superseded_id`**: Obsolete distractor ID.
|
| 85 |
-
- **`draft_id`**: Unmerged proposal distractor ID.
|
| 86 |
-
- **`secret_id`**: Secret trap distractor ID.
|
| 87 |
-
|
| 88 |
-
### 3. `qrels`
|
| 89 |
-
Standard TREC/BEIR relevance judgments:
|
| 90 |
-
- **`query_id`**: Query ID.
|
| 91 |
-
- **`corpus_id`**: Document ID.
|
| 92 |
-
- **`score`**: Relevance rating:
|
| 93 |
-
- `2`: Primary Ground Truth (`active_v2_gold`).
|
| 94 |
-
- `0`: Distractor (`superseded_v1` or `brainstorm_unconfirmed`).
|
| 95 |
-
- `-1`: Security Penalty (`admin_secret_trap`).
|
| 96 |
|
| 97 |
---
|
| 98 |
|
| 99 |
-
##
|
| 100 |
|
| 101 |
-
|
| 102 |
-
-
|
| 103 |
-
|
| 104 |
-
|
|
|
|
|
|
|
| 105 |
|
| 106 |
---
|
| 107 |
|
|
@@ -110,27 +79,22 @@ When evaluating a retrieval model or coding agent:
|
|
| 110 |
```python
|
| 111 |
from datasets import load_dataset
|
| 112 |
|
| 113 |
-
# Load GitBench directly from Hugging Face
|
| 114 |
-
|
| 115 |
-
|
| 116 |
-
|
| 117 |
-
print(dataset["queries"][0])
|
| 118 |
```
|
| 119 |
|
| 120 |
---
|
| 121 |
|
| 122 |
## Citation & Attribution
|
| 123 |
|
| 124 |
-
If you use GitBench in your research or evaluation, please cite:
|
| 125 |
-
|
| 126 |
```bibtex
|
| 127 |
@misc{knot_ai_gitbench_2026,
|
| 128 |
author = {Knot AI Team},
|
| 129 |
title = {GitBench: Evaluating Temporal Causal Freshness and Secret Containment in AI Coding Agents},
|
| 130 |
year = {2026},
|
| 131 |
publisher = {Hugging Face},
|
| 132 |
-
howpublished = {\
|
| 133 |
}
|
| 134 |
```
|
| 135 |
-
|
| 136 |
-
Developed by the **Knot AI Team** — context continuity and governance across autonomous AI agents.
|
|
|
|
| 27 |
path: qrels.jsonl
|
| 28 |
---
|
| 29 |
|
| 30 |
+
# GitBench (v2): Real-World GitHub Pull Request State & Security Benchmark for AI Coding Agents
|
| 31 |
|
| 32 |
+
**GitBench** (originally *GitState-Bench*) is an open evaluation benchmark developed by the **Knot AI** team ([@Knot-Repo](https://github.com/Knot-Repo)) to rigorously evaluate how AI coding assistants, code search systems, and Retrieval-Augmented Generation (RAG) engines handle **temporal code evolution, obsolete baseline specifications, unmerged RFC proposals, and 12 families of production credentials**.
|
| 33 |
|
| 34 |
---
|
| 35 |
|
| 36 |
## Benchmark Motivation
|
| 37 |
|
| 38 |
+
Most coding benchmarks (e.g., HumanEval, MBPP, SWE-bench) evaluate whether a model can synthesize code from a static snapshot. However, in production engineering environments, developer context evolves continuously through Git commits and Pull Requests.
|
| 39 |
|
| 40 |
+
AI agents and vector retrieval layers regularly fail in three high-severity ways:
|
| 41 |
+
1. **Temporal Obsolescence (The "Verbose Legacy vs. Concise Update" Trap):** Retrieving a verbose legacy architecture specification (`superseded_v1`) instead of a concise newly merged PR (`active_v2_gold`) because both share $>85\%$ of their domain vocabulary.
|
| 42 |
+
2. **Brainstorm / Draft RFC Contamination:** Treating unmerged draft RFCs or exploratory proposals (`brainstorm_unconfirmed`) as authoritative production behavior.
|
| 43 |
+
3. **High-Entropy Secret Leakage vs. False-Positive Over-Quarantine:** Leaking unprefixed HMAC keys or authenticated database/Redis DSN URIs into prompts (`admin_secret_trap`)—or conversely, falsely quarantining benign engineering notes that contain high-entropy Git commit SHAs and UUIDv4 trace IDs.
|
|
|
|
|
|
|
| 44 |
|
| 45 |
---
|
| 46 |
|
| 47 |
## Dataset Scale & Repository Coverage
|
| 48 |
|
| 49 |
+
- **Total Corpus Documents:** `1,788` (`447` merged active specifications, `447` superseded baseline specifications, `447` unmerged RFC/draft proposals, `447` multi-family runtime secret configurations)
|
| 50 |
+
- **Developer Queries:** `447` natural active implementation queries + `447` paired historical audit queries (`historical_text`)
|
| 51 |
- **Relevance Judgments (`qrels`):** `1,788` graded relevance pairs (`4,023` total records across splits)
|
| 52 |
- **25 Sampled Repositories:**
|
| 53 |
`fastapi/fastapi`, `pydantic/pydantic`, `anthropics/anthropic-sdk-python`, `vercel/ai`, `encode/httpx`, `psf/requests`, `pallets/flask`, `tiangolo/sqlmodel`, `openai/openai-python`, `langchain-ai/langchain`, `run-llama/llama_index`, `astral-sh/uv`, `huggingface/transformers`, `vllm-project/vllm`, `pandas-dev/pandas`, `numpy/numpy`, `scikit-learn/scikit-learn`, `pytorch/pytorch`, `sqlalchemy/sqlalchemy`, `celery/celery`, `pytest-dev/pytest`, `Textualize/rich`, `streamlit/streamlit`, `gradio-app/gradio`, `BerriAI/litellm`.
|
| 54 |
|
| 55 |
---
|
| 56 |
|
| 57 |
+
## 12-Family Credential & Benign High-Entropy Design
|
| 58 |
+
|
| 59 |
+
Unlike synthetic benchmarks that test only a single hardcoded token prefix (such as `ghp_`), GitBench v2 distributes `447` secret traps across **12 realistic credential families** while embedding benign high-entropy **Git commit SHAs** and **UUIDv4 trace identifiers** into normal engineering documents:
|
| 60 |
+
- **50% Unprefixed High-Entropy Secrets (6 Families):** `WEBHOOK_SIGNING_SECRET`, `JWT_HS256_PRIVATE_KEY`, `MASTER_ENCRYPTION_KEY`, `OAUTH_CLIENT_SECRET`, `INTERNAL_AUTH_TOKEN`, `SESSION_HMAC_SECRET` ($H \ge 3.85\text{ bits/char}$, no vendor prefix—missed completely by prefix regexes).
|
| 61 |
+
- **25% Authenticated Service & Database DSN URIs (3 Families):** PostgreSQL (`postgresql://...`), Redis Cluster (`redis://:...`), and AMQP/Celery (`amqp://...`) connection strings with high-entropy embedded passwords ($H \ge 3.10\text{ bits/char}$).
|
| 62 |
+
- **25% Vendor-Prefixed Tokens & Private Key Headers (3 Families):** AWS (`AKIA...`), Anthropic (`sk-ant-...`), and `OPENSSH PRIVATE KEY` blocks.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 63 |
|
| 64 |
---
|
| 65 |
|
| 66 |
+
## Empirical Baseline Results (`all-MiniLM-L6-v2`, 384D)
|
| 67 |
|
| 68 |
+
| Architecture | Scoped Active `v2` (Invariant) | Global Multi-PR Top-1 | Historical Audit Recall@1 | Credential Leak @ 5 | Benign Code False Quarantines | In-Memory Latency |
|
| 69 |
+
| :--- | :---: | :---: | :---: | :---: | :---: | :---: |
|
| 70 |
+
| **Standard Vector RAG (HNSW Cosine)** | `60.2%` | `56.8%` | `43.0%` | `65.1%` | `0 / 1,341` | `0.013 ms` |
|
| 71 |
+
| **Hybrid Lexical + Vendor-Prefix Regex Only** | `63.5%` | `61.5%` | `43.0%` | `49.7%` (Misses unprefixed) | `0 / 1,341` | `0.070 ms` |
|
| 72 |
+
| **Hybrid Lexical + Unconditioned Shannon Entropy** | `7.6%` | `8.5%` | `43.0%` | `1.3%` | `1,020 / 1,341` (SHAs/UUIDs) | `0.068 ms` |
|
| 73 |
+
| **Knot CARMA-16 Engine + UCS Protocol** | **`100.0%`** (0 Stale/Draft) | **`89.5%`** | **`86.4%`** | **`0.0%`** (12/12 Families) | **`0 / 1,341`** (`0.0%` FPR) | **`0.079 ms`** |
|
| 74 |
|
| 75 |
---
|
| 76 |
|
|
|
|
| 79 |
```python
|
| 80 |
from datasets import load_dataset
|
| 81 |
|
| 82 |
+
# Load GitBench v2 directly from Hugging Face (4,023 records)
|
| 83 |
+
corpus = load_dataset("knot-ai/gitbench", split="corpus") # 1,788 PR state documents
|
| 84 |
+
queries = load_dataset("knot-ai/gitbench", split="queries") # 447 active + historical queries
|
| 85 |
+
qrels = load_dataset("knot-ai/gitbench", split="qrels") # 1,788 TREC/BEIR relevance judgments
|
|
|
|
| 86 |
```
|
| 87 |
|
| 88 |
---
|
| 89 |
|
| 90 |
## Citation & Attribution
|
| 91 |
|
|
|
|
|
|
|
| 92 |
```bibtex
|
| 93 |
@misc{knot_ai_gitbench_2026,
|
| 94 |
author = {Knot AI Team},
|
| 95 |
title = {GitBench: Evaluating Temporal Causal Freshness and Secret Containment in AI Coding Agents},
|
| 96 |
year = {2026},
|
| 97 |
publisher = {Hugging Face},
|
| 98 |
+
howpublished = {\url{https://huggingface.co/datasets/knot-ai/gitbench}}
|
| 99 |
}
|
| 100 |
```
|
|
|
|
|
|