From d9c069ed86d70fd8bc5e4940d0c837030c59031c Mon Sep 17 00:00:00 2001 From: ModelHub XC Date: Mon, 31 Aug 2026 02:40:16 +0800 Subject: [PATCH] =?UTF-8?q?=E5=88=9D=E5=A7=8B=E5=8C=96=E9=A1=B9=E7=9B=AE?= =?UTF-8?q?=EF=BC=8C=E7=94=B1ModelHub=20XC=E7=A4=BE=E5=8C=BA=E6=8F=90?= =?UTF-8?q?=E4=BE=9B=E6=A8=A1=E5=9E=8B?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Model: gabriellarson/Foundation-Sec-8B-Instruct-GGUF Source: Original Platform --- .gitattributes | 59 ++++++ Foundation-Sec-8B-Instruct-F16.gguf | 3 + Foundation-Sec-8B-Instruct-IQ2_M.gguf | 3 + Foundation-Sec-8B-Instruct-IQ2_S.gguf | 3 + Foundation-Sec-8B-Instruct-IQ2_XS.gguf | 3 + Foundation-Sec-8B-Instruct-IQ2_XXS.gguf | 3 + Foundation-Sec-8B-Instruct-IQ3_M.gguf | 3 + Foundation-Sec-8B-Instruct-IQ3_S.gguf | 3 + Foundation-Sec-8B-Instruct-IQ3_XS.gguf | 3 + Foundation-Sec-8B-Instruct-IQ3_XXS.gguf | 3 + Foundation-Sec-8B-Instruct-IQ4_NL.gguf | 3 + Foundation-Sec-8B-Instruct-IQ4_XS.gguf | 3 + Foundation-Sec-8B-Instruct-Q2_K.gguf | 3 + Foundation-Sec-8B-Instruct-Q2_K_S.gguf | 3 + Foundation-Sec-8B-Instruct-Q3_K_L.gguf | 3 + Foundation-Sec-8B-Instruct-Q3_K_M.gguf | 3 + Foundation-Sec-8B-Instruct-Q3_K_S.gguf | 3 + Foundation-Sec-8B-Instruct-Q4_0.gguf | 3 + Foundation-Sec-8B-Instruct-Q4_K_M.gguf | 3 + Foundation-Sec-8B-Instruct-Q4_K_S.gguf | 3 + Foundation-Sec-8B-Instruct-Q5_0.gguf | 3 + Foundation-Sec-8B-Instruct-Q5_K_M.gguf | 3 + Foundation-Sec-8B-Instruct-Q5_K_S.gguf | 3 + Foundation-Sec-8B-Instruct-Q6_K.gguf | 3 + Foundation-Sec-8B-Instruct-Q8_0.gguf | 3 + README.md | 227 ++++++++++++++++++++++++ 26 files changed, 358 insertions(+) create mode 100644 .gitattributes create mode 100644 Foundation-Sec-8B-Instruct-F16.gguf create mode 100644 Foundation-Sec-8B-Instruct-IQ2_M.gguf create mode 100644 Foundation-Sec-8B-Instruct-IQ2_S.gguf create mode 100644 Foundation-Sec-8B-Instruct-IQ2_XS.gguf create mode 100644 Foundation-Sec-8B-Instruct-IQ2_XXS.gguf create mode 100644 Foundation-Sec-8B-Instruct-IQ3_M.gguf create mode 100644 Foundation-Sec-8B-Instruct-IQ3_S.gguf create mode 100644 Foundation-Sec-8B-Instruct-IQ3_XS.gguf create mode 100644 Foundation-Sec-8B-Instruct-IQ3_XXS.gguf create mode 100644 Foundation-Sec-8B-Instruct-IQ4_NL.gguf create mode 100644 Foundation-Sec-8B-Instruct-IQ4_XS.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q2_K.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q2_K_S.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q3_K_L.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q3_K_M.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q3_K_S.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q4_0.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q4_K_M.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q4_K_S.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q5_0.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q5_K_M.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q5_K_S.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q6_K.gguf create mode 100644 Foundation-Sec-8B-Instruct-Q8_0.gguf create mode 100644 README.md diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..5d757bb --- /dev/null +++ b/.gitattributes @@ -0,0 +1,59 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-F16.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-IQ2_M.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-IQ2_S.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-IQ2_XS.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-IQ2_XXS.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-IQ3_M.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-IQ3_S.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-IQ3_XS.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-IQ3_XXS.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-IQ4_NL.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-IQ4_XS.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q2_K.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q2_K_S.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q3_K_L.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q3_K_M.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q3_K_S.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q4_K_S.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q5_0.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q5_K_S.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text +Foundation-Sec-8B-Instruct-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text diff --git a/Foundation-Sec-8B-Instruct-F16.gguf b/Foundation-Sec-8B-Instruct-F16.gguf new file mode 100644 index 0000000..3d59fb8 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-F16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:017e447456c88ead5e454fd170a28d76af87f2126336e8a494746dc904493fd8 +size 16070994944 diff --git a/Foundation-Sec-8B-Instruct-IQ2_M.gguf b/Foundation-Sec-8B-Instruct-IQ2_M.gguf new file mode 100644 index 0000000..0c62963 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-IQ2_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bf10ec2f9dddb82e54b6fc2c5cb89c06a65bf9936e5d3b5803135718794a13ec +size 2948873472 diff --git a/Foundation-Sec-8B-Instruct-IQ2_S.gguf b/Foundation-Sec-8B-Instruct-IQ2_S.gguf new file mode 100644 index 0000000..159ba91 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-IQ2_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8a7ae01ec36a0ae179a9cc47a61ed6e1d20d741e2af077d71bc009755320849a +size 2759081216 diff --git a/Foundation-Sec-8B-Instruct-IQ2_XS.gguf b/Foundation-Sec-8B-Instruct-IQ2_XS.gguf new file mode 100644 index 0000000..07b94ca --- /dev/null +++ b/Foundation-Sec-8B-Instruct-IQ2_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d4d1eafe2329a23da6666d92ca7848202f9e19a1bc5cfc4e01a6d7a6e61b5ffc +size 2606320896 diff --git a/Foundation-Sec-8B-Instruct-IQ2_XXS.gguf b/Foundation-Sec-8B-Instruct-IQ2_XXS.gguf new file mode 100644 index 0000000..a87f1d5 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-IQ2_XXS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:47a1f5d7103eca58128418e605739394df4ced7e27aa4800ebcbc5524775af83 +size 2399751424 diff --git a/Foundation-Sec-8B-Instruct-IQ3_M.gguf b/Foundation-Sec-8B-Instruct-IQ3_M.gguf new file mode 100644 index 0000000..9d53407 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-IQ3_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c80a8a76283a290ed6aca6c07f65086e38b721048e32c12896028dc2ec12d8cb +size 3785485568 diff --git a/Foundation-Sec-8B-Instruct-IQ3_S.gguf b/Foundation-Sec-8B-Instruct-IQ3_S.gguf new file mode 100644 index 0000000..6cde52c --- /dev/null +++ b/Foundation-Sec-8B-Instruct-IQ3_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3ee8e090863b9224b9e86ddbbd1c57f41c07f3d417ca082e68f5b199f7e18d32 +size 3682987264 diff --git a/Foundation-Sec-8B-Instruct-IQ3_XS.gguf b/Foundation-Sec-8B-Instruct-IQ3_XS.gguf new file mode 100644 index 0000000..1ef8de4 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-IQ3_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2f7d24984bdf63fbee192061f60ea1074fda965d910791c5041e161b3289fffe +size 3519409408 diff --git a/Foundation-Sec-8B-Instruct-IQ3_XXS.gguf b/Foundation-Sec-8B-Instruct-IQ3_XXS.gguf new file mode 100644 index 0000000..eba6724 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-IQ3_XXS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2872e7240a62bd697569167f57172cab8d98d3fc5cdb53de74ab649ec707a8ad +size 3275504896 diff --git a/Foundation-Sec-8B-Instruct-IQ4_NL.gguf b/Foundation-Sec-8B-Instruct-IQ4_NL.gguf new file mode 100644 index 0000000..2050979 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-IQ4_NL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:59e535e9096f187ab5c1fe52e3b66ca035c77dc77f8897ceaf2e60a11fab71af +size 4678720768 diff --git a/Foundation-Sec-8B-Instruct-IQ4_XS.gguf b/Foundation-Sec-8B-Instruct-IQ4_XS.gguf new file mode 100644 index 0000000..90fdab5 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-IQ4_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3ffda7f9c9952e41e6b70d211be09facea176d0e9cff88f8d042ac5e587a1962 +size 4448378112 diff --git a/Foundation-Sec-8B-Instruct-Q2_K.gguf b/Foundation-Sec-8B-Instruct-Q2_K.gguf new file mode 100644 index 0000000..440941c --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7c69cd7f8555742e0e27158ccfcb763eb21dbac8cb9037af53ae52db6a8d27c5 +size 3179740416 diff --git a/Foundation-Sec-8B-Instruct-Q2_K_S.gguf b/Foundation-Sec-8B-Instruct-Q2_K_S.gguf new file mode 100644 index 0000000..5fe84fc --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q2_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2e41b484d2cf6c6acccbe27422a03b7391dc1423e79a23b60f3c166ef1eda34b +size 2989423872 diff --git a/Foundation-Sec-8B-Instruct-Q3_K_L.gguf b/Foundation-Sec-8B-Instruct-Q3_K_L.gguf new file mode 100644 index 0000000..8136127 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q3_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:02aee7e492596c3b0b76095e8a92fd730458748c829adb2d1a78ef2c839198dc +size 4322618624 diff --git a/Foundation-Sec-8B-Instruct-Q3_K_M.gguf b/Foundation-Sec-8B-Instruct-Q3_K_M.gguf new file mode 100644 index 0000000..c5711e0 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q3_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8f483a31ef71a14ef0c8059a7aff8c4cab4f59c59fdd4d901d3bedd612ce034c +size 4019580160 diff --git a/Foundation-Sec-8B-Instruct-Q3_K_S.gguf b/Foundation-Sec-8B-Instruct-Q3_K_S.gguf new file mode 100644 index 0000000..9e0c21d --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q3_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:59e8e0fba90ed49c8d2789c6210d621a75e444d4017a009ede3e280ce2fd8e39 +size 3665161472 diff --git a/Foundation-Sec-8B-Instruct-Q4_0.gguf b/Foundation-Sec-8B-Instruct-Q4_0.gguf new file mode 100644 index 0000000..012a45e --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:583e9bef8a3645fc695961196643357bab83294c6d740385f54a19a227053b40 +size 4676623616 diff --git a/Foundation-Sec-8B-Instruct-Q4_K_M.gguf b/Foundation-Sec-8B-Instruct-Q4_K_M.gguf new file mode 100644 index 0000000..34319ce --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3175e83aa66aefeeddbc17608e3e8a2f52827941c73ccbe14fc9d138c8370206 +size 4921466112 diff --git a/Foundation-Sec-8B-Instruct-Q4_K_S.gguf b/Foundation-Sec-8B-Instruct-Q4_K_S.gguf new file mode 100644 index 0000000..21bdb5f --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff5d4ad1174fb14b21e052d75df4f78e30e5b7299272377eec0a08a2b4e04033 +size 4693400832 diff --git a/Foundation-Sec-8B-Instruct-Q5_0.gguf b/Foundation-Sec-8B-Instruct-Q5_0.gguf new file mode 100644 index 0000000..b59c4ff --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q5_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9066575cb3ad94b2afb0eecd2d10e33fea9131145b8cd6bf031c9e1b9b5c0f1e +size 5614771456 diff --git a/Foundation-Sec-8B-Instruct-Q5_K_M.gguf b/Foundation-Sec-8B-Instruct-Q5_K_M.gguf new file mode 100644 index 0000000..c51a240 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b7107f65790f087415f8ffc984fbe899ea2c1be93785d604da1332acf22cc008 +size 5733784832 diff --git a/Foundation-Sec-8B-Instruct-Q5_K_S.gguf b/Foundation-Sec-8B-Instruct-Q5_K_S.gguf new file mode 100644 index 0000000..d12c4b8 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:db01543df4e6a037a49d5f452808f544337e7964883caf54438285af26302140 +size 5600091392 diff --git a/Foundation-Sec-8B-Instruct-Q6_K.gguf b/Foundation-Sec-8B-Instruct-Q6_K.gguf new file mode 100644 index 0000000..4d1fc02 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8b4a372dcce4a4365515a2dcf5c4e9a65768259bcc7a6736ab8df2ae85bfc095 +size 6596873472 diff --git a/Foundation-Sec-8B-Instruct-Q8_0.gguf b/Foundation-Sec-8B-Instruct-Q8_0.gguf new file mode 100644 index 0000000..2a13ec5 --- /dev/null +++ b/Foundation-Sec-8B-Instruct-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2137ffe55c1654069c808f9da461ef11f64fbda8e37eeef1c5f5b258f07bd89a +size 8541891840 diff --git a/README.md b/README.md new file mode 100644 index 0000000..4ac6893 --- /dev/null +++ b/README.md @@ -0,0 +1,227 @@ +--- +base_model: +- fdtn-ai/Foundation-Sec-8B-Instruct +language: +- en +library_name: transformers +license: other +pipeline_tag: text-generation +tags: +- security +- llama +--- +# Foundation-Sec-8B-Instruct - Model Card + +## Model Information + +Llama-3.1-FoundationAI-SecurityLLM-8B-Instruct (Foundation-Sec-8B-Instruct) is an open-weight, 8-billion parameter instruction-tuned language model specialized for cybersecurity applications. +It extends the Foundation-Sec-8B base model with instruction-following capabilities. +It leverages prior training to understand security concepts, terminology, and practices across multiple security domains. +Further instruction-tuning allows the model to interact with human users in a chat-like interface. +Foundation-Sec-8B-Instruct enables organizations to build AI-driven security tools that can be deployed locally, reducing dependency on cloud-based AI services while maintaining high performance on security-related tasks. + +- **Model Name:** Llama-3.1-FoundationAI-SecurityLLM-8B-Instruct (Foundation-Sec-8B-Instruct) +- **Model Developer:** Amin Karbasi and Research team at Foundation AI — Cisco +- **Model Card Contact:** For questions about the team, model usage, and future directions, contact [`karbasi@cisco.com`](mailto:karbasi@cisco.com). For technical questions about the model, please contact [`sajanaw@cisco.com`](mailto:sajanaw@cisco.com) and [`paulkass@cisco.com`](mailto:paulkass@cisco.com). +- **Model Release Date:** August 1st, 2025 +- **Supported Language(s):** English +- **Model Architecture:** Auto-regressive language model that uses an optimized transformer architecture (Meta Llama-3.1-8B backbone) +- **Training Objective:** Instruction following and alignment with human preferences +- **Training Data Status:** This is a static model trained on an offline dataset. Future versions of the tuned models will be released on updated data. +- **License:** See NOTICE.md + + +## Intended Use + +### Intended Use Cases + +Foundation-Sec-8B-Instruct is designed for security practitioners, researchers, and developers building AI-powered security workflows and applications. +Foundation-Sec-8B-Instruct is optimized for three core use case categories: + +- **SOC Acceleration**: Automating triage, summarization, case note generation, and evidence collection. +- **Proactive Threat Defense**: Simulating attacks, prioritizing vulnerabilities, mapping TTPs, and modeling attacker behavior. +- **Engineering Enablement**: Providing security assistance, validating configurations, assessing compliance evidence, and improving security posture. + +The model is intended for local deployment in environments prioritizing data security, regulatory compliance, and operational control. + +### Downstream Use + +Foundation-Sec-8B-Instruct can be used directly for security-related chat use cases. Example downstream applications include: + +- Summarization + - Summarizing detection playbooks and incident reports + - Consolidating fragmented analyst notes into structured case summaries +- Classification + - Mapping threats to MITRE ATT&CK techniques + - Prioritizing vulnerabilities based on contextual risk + - Classifying security-relevant emails and leaked file contents +- Named Entity Recognition + - Extracting compliance evidence from documents + - Building network behavior profiles from technical manuals +- Question & Answer + - Assisting SOC analysts with alert triage and investigation + - Responding to cloud security and software compliance queries +- Reasoning and Text Generation + - Generating red-team attack plans and threat models + - Predicting attacker next steps in active investigations + - Enriching vulnerability scan results with contextual insights + +For questions or assistance with fine-tuning Foundation-Sec-8B-Instruct, please reach out to the team. + +### Out-of-Scope Use + +The following uses are out-of-scope and are neither recommended nor intended use cases: + +1. **Generating harmful content** - The model should not be used to: + - Generate malware or other malicious code + - Create phishing content or social engineering scripts + - Develop attack plans targeting specific organizations + - Design exploitation techniques for vulnerabilities without legitimate security research purposes +2. **Critical security decisions without human oversight** - The model should not be used for: + - Autonomous security decision-making without human review + - Critical infrastructure protection without expert supervision + - Final determination of security compliance without human verification + - Autonomous vulnerability remediation without testing +3. **Legal or medical advice** - The model is not qualified to provide: + - Legal advice regarding security regulations, compliance requirements, or intellectual property disputes + - Legal advice regarding security issues that would reference legal statutes, precedents, or case law necessary to provide legal advice + - Medical advice regarding health impacts of security incidents +4. **Non-security use cases** - The model is specifically optimized for cybersecurity and may not perform as well on general tasks as models trained for broader applications. +5. **Violation of Laws or Regulations** - Any use that violates applicable laws or regulations. + +## How to Get Started with the Model + +Use the code below to get started with the model. + +```python +# Import the required libraries +import torch +from transformers import AutoTokenizer, AutoModelForCausalLM +# Load the model and tokenizer +tokenizer = AutoTokenizer.from_pretrained("fdtn-ai/Foundation-Sec-8B-Instruct") +model = AutoModelForCausalLM.from_pretrained("fdtn-ai/Foundation-Sec-8B-Instruct") +prompt = "CVE-2015-10011 is a vulnerability about OpenDNS OpenResolve improper log output neutralization. What is the corresponding CWE?" +messages = [ + {"role": "user", "content": prompt} +] +model_inputs = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) +inputs = tokenizer(model_inputs, return_tensors="pt", add_special_tokens=False) +output = model.generate(**inputs, temperature=0.1, max_new_tokens=250) +resp = tokenizer.batch_decode(output)[0] +print(resp.replace(model_inputs, "")) +``` + +## Training and Evaluation + +### Training Data + +Foundation-Sec-8B-Instruct was trained on a wide variety of public and proprietary question answer/pairs for general and security-specific instruction-following. + +**Data cutoff:** April 10th, 2025. + +A more detailed description of the methodology is available in the technical report. + +### Training Setup + +Foundation-Sec-8B-Instruct is based on the **Llama 3.1 8B** architecture. Training was performed on Cisco Foundation AI’s internal compute cluster. + +Key training details: + +- **Instruction fine-tuning** to follow human instructions +- **RLHF** to align model answers to human preferences +- **4096-token** sequence length +- **Optimizer:** AdamW + +A more detailed description of the methodology is available in the technical report. + +### Evaluation + +Foundation-Sec-8B-Instruct was benchmarked on cybersecurity and general reasoning tasks, using a standardized 0-shot instruction prompting setup (temperature = 0.3). + +| **Benchmark** | **Foundation-sec-8B** | **Llama 3.1 8B** | **GPT-4o-mini** | +| --- | --- | --- | --- | +| CTI-MCQA | 0.644 | 0.617 | 0.672 | +| CTI-RCM | 0.692 | 0.558 | 0.655 | +| CTI-VSP | 0.802 | 0.815 | 0.792 | +| IF-Eval | 0.811 | 0.791 | 0.834 | +| Alpaca Eval 2 | 35.453 | 24.477 | 52.720 | + +**Benchmark Overview:** + +- **CTI-MCQA:** 2,500 multiple-choice questions testing cybersecurity knowledge across frameworks like MITRE ATT&CK, NIST, GDPR, and threat intelligence best practices. +- **CTI-RCM:** 1,000 vulnerability root cause mapping examples linking CVEs to CWE categories, assessing deep understanding of security weaknesses. +- **CTI-VSP:** A set of 1,000 CVE descriptions where models predict the CVSS v3 Base metrics and compute the overall score, with performance measured by the average absolute difference from the true scores. +- **IF-Eval:** 541 instruction-following prompts designed for automated, reproducible assessment of LLM instruction-following capabilities. +- **Alpaca Eval 2:** 805 single-turn prompts auto-scored by GPT-4 Turbo against a GPT-4 Turbo reference, validated with 20,000 human preference votes, and closely matching ChatBot Arena results. + +**Key highlights:** + +- **+3 to +11 point gains** over Llama-3.1-8B-Instruct across security-specific benchmarks. +- **Exceptional Instruction-Following capabilities** exceeding that of Llama-3.1-8B-Instruct. +- **Competitive against small Frontier Models** such as GPT-4o-mini on instruction-following capabilities and cybersecurity tasks. + +For full benchmark details and evaluation methodology, please refer to the technical report. + +## Safety Alignment + +Standard best practices were followed to align the model with general safety values. +Despite the alignment, however, safe out-of-the-box performance cannot be guaranteed. +Our evaluations show that while the model can achieve reasonable safety performance out-of-the-box, LlamaGuard provides much better protection against malicious requests. +It is recommended to deploy this model with additional safeguards (such as LlamaGuard) and human oversight. + +| Model | HarmBench Performance | +|---|---| +| Llama-3.1-8b-Instruct | 72.43% | +| Foundation-Sec-8B-Instruct | 91.98% | +| **LlamaGuard** + Foundation-Sec-8B-Instruct | 99.25% | + + +## Limitations + +Foundation-Sec-8B-Instruct has several limitations that users should be aware of: + +1. **Domain-specific knowledge limitations**: + - Foundation-Sec-8B-Instruct may not be familiar with recent vulnerabilities, exploits, or novel attack vectors or security technologies released after its training cutoff date + - Knowledge of specialized or proprietary security systems or tools may be limited +2. **Potential biases**: + - The model may reflect biases present in security literature and documentation + - The model may be trained on known attack patterns and have difficulty recognizing novel attack vectors + - Security practices and recommendations may be biased toward certain technological ecosystems + - Geographic and cultural biases in security approaches may be present +3. **Security risks**: + - The model cannot verify the identity or intentions of users + - Adversarial prompting techniques might potentially bypass safety mechanisms + - The model may unintentionally provide information that could be misused if proper prompting guardrails are not implemented +4. **Contextual blindness:** + - The model may struggle to understand the complex interrelationships between systems, users, and data in order to provide accurate context. +5. **Technical limitations**: + - Performance varies based on how security concepts are described in prompts + - May not fully understand complex, multi-step security scenarios without clear explanation + - Cannot access external systems or actively scan environments + - Cannot independently verify factual accuracy of its outputs +6. **Ethical considerations**: + - Dual-use nature of security knowledge requires careful consideration of appropriate use cases + +### Recommendations + +To address the limitations of Foundation-Sec-8B-Instruct, we recommend: + +1. **Human oversight**: + - Always have qualified security professionals review model outputs before implementation + - Use the model as an assistive tool rather than a replacement for expert human judgment + - Implement a human-in-the-loop approach for security-critical applications +2. **System design safeguards**: + - Implement additional validation layers for applications built with this model + - Consider architectural constraints that limit the model's ability to perform potentially harmful actions (excessive agency) + - Deploy the model in environments with appropriate access controls +3. **Prompt engineering**: + - Use carefully designed prompts that encourage ethical security practices + - Include explicit instructions regarding responsible disclosure and ethical hacking principles + - Structure interactions to minimize the risk of inadvertently harmful outputs +4. **Knowledge supplementation**: + - Supplement the model with up-to-date security feeds and databases + - Implement retrieval-augmented generation for current threat intelligence sources +5. **Usage policies**: + - Develop and enforce clear acceptable use policies for applications using this model + - Implement monitoring and auditing for high-risk applications + - Create documentation for end users about the model's limitations \ No newline at end of file