<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Dor Cohen</title><description>Software engineer and researcher — ML safety, security, and production systems.</description><link>https://dorcohen.dev/</link><item><title>CNF Entropy Approximation for SAT Model Counting</title><link>https://dorcohen.dev/posts/cnf-entropy-approximation/</link><guid isPermaLink="true">https://dorcohen.dev/posts/cnf-entropy-approximation/</guid><description>How I used uniform solution sampling to approximate Shannon entropy for CNF formulas, applied to a SAT model counter.</description><pubDate>Mon, 19 Jun 2017 00:00:00 GMT</pubDate></item><item><title>Estimating Causal Effects with Multiple Treatments</title><link>https://dorcohen.dev/posts/multiple-treatment-causal-effect/</link><guid isPermaLink="true">https://dorcohen.dev/posts/multiple-treatment-causal-effect/</guid><description>Comparing covariate adjustment and matching methods for estimating causal effects when there are more than two treatment groups.</description><pubDate>Mon, 01 Feb 2021 00:00:00 GMT</pubDate></item><item><title>What Can AI Actually Hack? Inside Frontier Labs&apos; Tests</title><link>https://dorcohen.dev/posts/what-can-ai-actually-hack/</link><guid isPermaLink="true">https://dorcohen.dev/posts/what-can-ai-actually-hack/</guid><description>What each cyber task really is, where it runs, and how far today&apos;s models get.</description><pubDate>Sun, 08 Nov 2026 00:00:00 GMT</pubDate></item></channel></rss>