From e2ca8daee1fdc8586851943350128f9bed8ffca1 Mon Sep 17 00:00:00 2001
From: Prashant Dixit <54981696+PrashantDixit0@users.noreply.github.com>
Date: Wed, 2 Oct 2024 21:15:24 +0530
Subject: [PATCH] docs: saleforce's sfr rag (#1717)

This PR adds Salesforce's newly released SFR RAG
---
 docs/mkdocs.yml         |  5 ++++-
 docs/src/rag/sfr_rag.md | 17 +++++++++++++++++
 2 files changed, 21 insertions(+), 1 deletion(-)
 create mode 100644 docs/src/rag/sfr_rag.md

diff --git a/docs/mkdocs.yml b/docs/mkdocs.yml
index 5c91577f..5abf132e 100644
--- a/docs/mkdocs.yml
+++ b/docs/mkdocs.yml
@@ -120,6 +120,7 @@ nav:
               - Graph RAG: rag/graph_rag.md
               - Self RAG: rag/self_rag.md
               - Adaptive RAG: rag/adaptive_rag.md
+              - SFR RAG: rag/sfr_rag.md
               - Advanced Techniques:
                 - HyDE: rag/advanced_techniques/hyde.md
                 - FLARE: rag/advanced_techniques/flare.md
@@ -247,6 +248,7 @@ nav:
           - Graph RAG: rag/graph_rag.md
           - Self RAG: rag/self_rag.md
           - Adaptive RAG: rag/adaptive_rag.md
+          - SFR RAG: rag/sfr_rag.md
           - Advanced Techniques:
             - HyDE: rag/advanced_techniques/hyde.md
             - FLARE: rag/advanced_techniques/flare.md
@@ -362,4 +364,5 @@ extra:
     - icon: fontawesome/brands/x-twitter
       link: https://twitter.com/lancedb
     - icon: fontawesome/brands/linkedin
-      link: https://www.linkedin.com/company/lancedb
+      link: https://www.linkedin.com/company/lancedb      
+      
\ No newline at end of file
diff --git a/docs/src/rag/sfr_rag.md b/docs/src/rag/sfr_rag.md
new file mode 100644
index 00000000..9f063575
--- /dev/null
+++ b/docs/src/rag/sfr_rag.md
@@ -0,0 +1,17 @@
+**SFR RAG 📑**
+====================================================================
+Salesforce AI Research introduces SFR-RAG, a 9-billion-parameter language model trained with a significant emphasis on reliable, precise, and faithful contextual generation abilities specific to real-world RAG use cases and relevant agentic tasks. They include precise factual knowledge extraction, distinguishing relevant against distracting contexts, citing appropriate sources along with answers, producing complex and multi-hop reasoning over multiple contexts, consistent format following, as well as refraining from hallucination over unanswerable queries.
+
+**[Offical Implementation](https://github.com/SalesforceAIResearch/SFR-RAG)**
+
+<figure markdown="span">
+  ![agent-based-rag](https://raw.githubusercontent.com/lancedb/assets/main/docs/assets/rag/salesforce_contextbench.png)
+  <figcaption>Average Scores in ContextualBench: <a href="https://blog.salesforceairesearch.com/sfr-rag/">Source</a>
+  </figcaption>
+</figure>
+
+To reliably evaluate LLMs in contextual question-answering for RAG, Saleforce introduced [ContextualBench](https://huggingface.co/datasets/Salesforce/ContextualBench?ref=blog.salesforceairesearch.com), featuring 7 benchmarks like [HotpotQA](https://arxiv.org/abs/1809.09600?ref=blog.salesforceairesearch.com) and [2WikiHopQA](https://www.aclweb.org/anthology/2020.coling-main.580/?ref=blog.salesforceairesearch.com) with consistent setups. 
+
+SFR-RAG outperforms GPT-4o, achieving state-of-the-art results in 3 out of 7 benchmarks, and significantly surpasses Command-R+ while using 10 times fewer parameters. It also excels at handling context, even when facts are altered or conflicting.
+
+[Saleforce AI Research Blog](https://blog.salesforceairesearch.com/sfr-rag/)