Unverified50% confidenceFactExact time
SafeTulu is a version of Tulu 3 built on SimpleLoRIF with the 20% of samples most aligned with the gradients of unsafe responses removed.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/28/2026
First Seen
Valid until: 1/26/2027
Sources
Related Entities
Related Claims
UnverifiedThe SafeTulu dataset has been made public, and a Hugging Face Space has also been released, allowing researchers to build custom versions of SafeTulu, for example by adjusting the removal ratio or filtering only a specific subset.55% similarVerifiedSafeTensors uses a simple header-plus-data structure that contains no executable code, eliminating the security risk of arbitrary code execution during deserialization present in the traditional pickle format.54% similarUnverified0.65.4 is a security backport for the stable branch.53% similarUnverifiedSturdybench designed three test cases (notool-001, missing-001, unsafe-006) in a free open-source repository, covering respectively three scenarios: unnecessary tool calls, reckless actions when required parameters are missing, and prompt injection attacks.52% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1105870API
curl https://kongchang.com/api/v1/knowledge/claims/1105870MCP
get_claim(id=1105870)