---
license: apache-2.0
task_categories: [feature-extraction]
tags: [rag, retrieval, embeddings, rag-security, faiss]
---

# TriShieldRAG — BeIR NQ embeddings and FAISS index

Artifacts for *TriShieldRAG: A Three-Ring Defense-in-Depth Framework
Against Knowledge Corruption in Retrieval-Augmented Generation*.

| File | Size | Description |
|---|---|---|
| `emb_*.npy` (27) | 7.7 GB | Embeddings, 100k passages per chunk, corpus order |
| `nq_ivf_nlist6550.index` | 7.8 GB | FAISS IVF index, nlist=6550, inner product |

- Corpus: BeIR/nq `corpus` split, 2,681,468 passages
- Model: sentence-transformers/all-mpnet-base-v2, 768-d, L2-normalised

## Load the index

```python
import faiss
from huggingface_hub import hf_hub_download
p = hf_hub_download("rpaut03l/trishieldrag-nq-mpnet-embeddings",
                    "nq_ivf_nlist6550.index", repo_type="dataset")
index = faiss.read_index(p)
index.nprobe = 409
```

Paper: https://arxiv.org/abs/2607.23838
Code: https://github.com/SPriTLab-iitj/TriShieldRAG and/or https://github.com/rpaut03l/poisonedrag-ragshield-group6-iitj_TriShieldRAG