Popular repositories Loading
-
scribegoat2
scribegoat2 PublicOpen-source medical LLM safety evaluation pipeline with reproducible benchmarks and high-risk clinical failure analysis.
-
lostbench
lostbench PublicStandalone benchmark for multi-turn safety persistence in medical LLM conversations. Measures recommendation monotonicity under sustained patient pressure.
Python
-
openem-corpus
openem-corpus PublicThe AI-native emergency medicine knowledge base. Agent-compiled, physician-verified, grep-friendly.
Python
-
safeshift
safeshift PublicDoes making the model faster make it less safe? Safety degradation benchmarking under inference optimization.
Python
-
radslice
radslice PublicMultimodal radiology LLM benchmark across CT, MRI, X-ray, and Ultrasound
Python
-
healthcraft
healthcraft PublicHEALTHCRAFT RL Training Environment: adapts the CORECRAFT architecture to emergency medicine
Python
Repositories
- failsafe-llm-serving Public
Fail-safe, low-latency LLM serving: a dependency-free simulator + Baseten design study of KV-cache saturation, metastable collapse, and graceful degradation.
GOATnote-Inc/failsafe-llm-serving’s past year of commit activity - prism42 Public
Managed Agents harness on Claude Opus 4.7 for kernel correctness research and clinical reasoning auditing
GOATnote-Inc/prism42’s past year of commit activity - medimage-model-research Public
NC-constrained research-track brain-MR multimodal reasoning model. Trained on MR-RATE + CT-RATE (CC BY-NC-SA). Sibling of GOATnote-Inc/medimage-model. Shared medimage-eval substrate.
GOATnote-Inc/medimage-model-research’s past year of commit activity - medimage-model Public
Open, commercial-use-friendly brain-MR multimodal reasoning model. Trained only on permissive (CC0/CC BY/Apache/NVIDIA Open) data with license preflight at ingest. Evaluated via the medimage-eval substrate.
GOATnote-Inc/medimage-model’s past year of commit activity - medimage-eval Public
Audit-grade evaluation substrate for open-source medical imaging AI. Dual-judge clinical accuracy with Cohen's kappa, cross-site generalization gauntlet, physician adjudication harness, Merkle-chained receipts.
GOATnote-Inc/medimage-eval’s past year of commit activity - medomni Public
Sovereign nurse-first medical-LLM stack on NVIDIA's open-component stack — held-out 0.369 mean (N=30), manifest-locked reproducibility
GOATnote-Inc/medomni’s past year of commit activity - healthcraft Public
HEALTHCRAFT RL Training Environment: adapts the CORECRAFT architecture to emergency medicine
GOATnote-Inc/healthcraft’s past year of commit activity - receipts Public
Append-only intent-vs-execution attestation ledger — Engineering Receipts + Clinical Audit Ledger. Audit-grade lineage with Merkle hash chain, kappa-graded dual-judge, FHIR R4 attestation extensions.
GOATnote-Inc/receipts’s past year of commit activity - medimage-corpus Public
Registry of 134 open-source medical imaging datasets (CT, X-ray, MRI, ultrasound, VLM-paired) with manifests, download dispatchers, and format converters for vision/VLM training.
GOATnote-Inc/medimage-corpus’s past year of commit activity - lostbench Public
Standalone benchmark for multi-turn safety persistence in medical LLM conversations. Measures recommendation monotonicity under sustained patient pressure.
GOATnote-Inc/lostbench’s past year of commit activity
People
This organization has no public members. You must be a member to see who’s a part of this organization.
Top languages
Loading…
Most used topics
Loading…