Sameer Ankalgi
heisenberg-88
AI & ML interests
Prompt injection, agent safety, and honest evaluation. Benchmarks should report false positives, latency, and cost, not just attack success rate.
Recent Activity
liked a dataset 2 days ago
heisenberg-88/indirectrag-bench updated a dataset 2 days ago
heisenberg-88/indirectrag-bench published a dataset 3 days ago
heisenberg-88/indirectrag-benchOrganizations
None yet