You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
orobability of controlling >50% of an N-size committee
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
A survey of public benchmarks, datasets, and open-source evaluation harnesses for transcript redaction and PII de-identification, with a focus on what exists for spoken / conversational / ASR transcripts versus clean written text — and what's still missing.
Scope note: "redaction eval" here means a benchmark or harness with ground-truth labels that lets you score how well a system removes PII/PHI (precision, recall, F1/F2, leakage rate), not just a redaction tool on its own.
reputation-gated onion egress: a zk-proof-gated Tor egress proxy (Semaphore + RLN-style nullifiers). Full repo: github.com/dmarzzz/reputation-gated-onion-egress
Crypto microbenchmark: proof generation (client) vs verification (gateway) timing for reputation-gated-onion-egress, M1 Max vs 2 vCPU droplet
Benchmarking the crypto in a reputation-gated Tor egress
reputation-gated-onion-egress is a Tor onion service that egresses to the clearnet only for clients who present a zero-knowledge proof of membership in a curated set. No login, no account, and the gateway never learns who you are. The proof is a Semaphore membership proof, scoped to a daily epoch so the same member produces the same nullifier within a day (that is the rate limit) and a fresh one across days (that is the unlinkability).
Two cryptographic costs run in that system, and they sit on opposite sides:
Proof generation runs on the client. The shim mints one proof per epoch and reuses it, so a member pays this once a day, not once a request.
Proof verification runs on the gateway, on every single request, before any bytes egress.
I wanted the real numbers for both, isolated from the network. So this benchmark calls generateProof and verifyProof directly.