Anthony Hughes
@antjhughes.bsky.social
PhD Student at University of Sheffield. Researching privacy in M/LMs.
Findings:
❓ Privacy-preservation at inference-time is really underexplored! 🔍 LMs struggle to prevent PII leakage in their summaries. 👩⚖️ Human evaluations reveal privacy risks that metrics may overlook.
Paper w/ @naletras.bsky.social and Ning Ma
Cc. @sltcdt.bsky.social
❓ Privacy-preservation at inference-time is really underexplored! 🔍 LMs struggle to prevent PII leakage in their summaries. 👩⚖️ Human evaluations reveal privacy risks that metrics may overlook.
Paper w/ @naletras.bsky.social and Ning Ma
Cc. @sltcdt.bsky.social
December 22, 2024 at 9:41 PM
Findings:
❓ Privacy-preservation at inference-time is really underexplored! 🔍 LMs struggle to prevent PII leakage in their summaries. 👩⚖️ Human evaluations reveal privacy risks that metrics may overlook.
Paper w/ @naletras.bsky.social and Ning Ma
Cc. @sltcdt.bsky.social
❓ Privacy-preservation at inference-time is really underexplored! 🔍 LMs struggle to prevent PII leakage in their summaries. 👩⚖️ Human evaluations reveal privacy risks that metrics may overlook.
Paper w/ @naletras.bsky.social and Ning Ma
Cc. @sltcdt.bsky.social
We tested across 5 LMs (both open & closed) and 3 domains. We analyzed both prompting and fine-tuning techniques to guide LMs toward safer summaries. Summarization datasets from medicine, legal, and general domains were used to measure how much PII leaks.
December 22, 2024 at 9:41 PM
We tested across 5 LMs (both open & closed) and 3 domains. We analyzed both prompting and fine-tuning techniques to guide LMs toward safer summaries. Summarization datasets from medicine, legal, and general domains were used to measure how much PII leaks.