Fingerprinting Leaderboard benchmarks model provenance as copying takes minutes vs millions to train
omarsar0 · x · 2026-10-14
Sentient launched the Fingerprinting Leaderboard, an open benchmark for model fingerprinting methods that embed provenance in weights rather than paperwork. Models cost millions to train but can be copied in minutes, and fine-tuning or distillation can erase their origins.
The leaderboard scores each method on three axes:
- How much model quality it preserves
- How well it survives removal attempts
- How reliably the owner can detect it
Omar Khattab-style curator omarsar urges anyone training or fine-tuning their own models to adopt fingerprinting to prove ownership once weights are released.
More from Research
- Snorkel expands Open Benchmarks Grants 10x to $30M, launches Red Team and fellowship — ajratner · 2026-10-14
- PORTAL: training-free projector init skips multimodal alignment with 135x fewer FLOPs — anshulkundaje · 2026-10-14
- Deep learning maps low-affinity pioneer binding sites that cooperate to open chromatin — anshulkundaje · 2026-10-14
- Agent team study: 84% of messages still go to the lead, workers rarely talk to each other — nouhadziri · 2026-10-14
- Sol lead + 16 cheap Luna planners nears all-Sol team results at 70% less cost — nouhadziri · 2026-10-14
- Delegate Only Planning: Agent Team Matches Full Delegation at Half the Cost — nouhadziri · 2026-10-14