<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>S. B. Cohen | TSG Lab – Technical Safety &amp; Governance Lab</title><link>https://tsglab.github.io/author/s.-b.-cohen/</link><atom:link href="https://tsglab.github.io/author/s.-b.-cohen/index.xml" rel="self" type="application/rss+xml"/><description>S. B. Cohen</description><generator>Hugo Blox Builder (https://hugoblox.com)</generator><language>en-us</language><lastBuildDate>Sun, 01 Mar 2026 00:00:00 +0000</lastBuildDate><image><url>https://tsglab.github.io/media/logo.svg</url><title>S. B. Cohen</title><link>https://tsglab.github.io/author/s.-b.-cohen/</link></image><item><title>Old Habits Die Hard: How Conversational History Geometrically Traps LLMs</title><link>https://tsglab.github.io/publication/old-habits-die-hard-conversational-history/</link><pubDate>Sun, 01 Mar 2026 00:00:00 +0000</pubDate><guid>https://tsglab.github.io/publication/old-habits-die-hard-conversational-history/</guid><description/></item><item><title>PoisonBench: Assessing Large Language Model Vulnerability to Data Poisoning</title><link>https://tsglab.github.io/publication/poisonbench/</link><pubDate>Tue, 01 Jul 2025 00:00:00 +0000</pubDate><guid>https://tsglab.github.io/publication/poisonbench/</guid><description/></item><item><title>Interpreting Context Look-ups in Transformers: Investigating Attention-MLP Interactions</title><link>https://tsglab.github.io/publication/attention-mlp-interactions/</link><pubDate>Fri, 01 Nov 2024 00:00:00 +0000</pubDate><guid>https://tsglab.github.io/publication/attention-mlp-interactions/</guid><description/></item><item><title>Large Language Models Relearn Removed Concepts</title><link>https://tsglab.github.io/publication/llms-relearn-removed-concepts/</link><pubDate>Thu, 01 Aug 2024 00:00:00 +0000</pubDate><guid>https://tsglab.github.io/publication/llms-relearn-removed-concepts/</guid><description/></item><item><title>The Scaling Behavior of Large Language Models</title><link>https://tsglab.github.io/publication/scaling-behavior-llms/</link><pubDate>Mon, 01 Jul 2024 00:00:00 +0000</pubDate><guid>https://tsglab.github.io/publication/scaling-behavior-llms/</guid><description/></item><item><title>What Does GPT Store in Its MLP Weights? A Case Study of Long-Range Dependencies</title><link>https://tsglab.github.io/publication/gpt-mlp-weights-long-range/</link><pubDate>Mon, 01 Jan 2024 00:00:00 +0000</pubDate><guid>https://tsglab.github.io/publication/gpt-mlp-weights-long-range/</guid><description/></item><item><title>The Larger They Are, the Harder They Fail: Language Models Do Not Recognize Identifier Swaps in Python</title><link>https://tsglab.github.io/publication/identifier-swaps-python/</link><pubDate>Sat, 01 Jul 2023 00:00:00 +0000</pubDate><guid>https://tsglab.github.io/publication/identifier-swaps-python/</guid><description/></item></channel></rss>