{"ID":2826451,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.18546","arxiv_id":"2512.18546","title":"LLMs on Drugs: Language Models Are Few-Shot Consumers","abstract":"Large language models (LLMs) are sensitive to the personas imposed on them at inference time, yet prompt-level \"drug\" interventions have never been benchmarked rigorously. We present the first controlled study of psychoactive framings on GPT-5-mini using ARC-Challenge. Four single-sentence prompts -- LSD, cocaine, alcohol, and cannabis -- are compared against a sober control across 100 validation items per condition, with deterministic decoding, full logging, Wilson confidence intervals, and Fisher exact tests. Control accuracy is 0.45; alcohol collapses to 0.10 (p = 3.2e-8), cocaine to 0.21 (p = 4.9e-4), LSD to 0.19 (p = 1.3e-4), and cannabis to 0.30 (p = 0.041), largely because persona prompts disrupt the mandated \"Answer: \u003cLETTER\u003e\" template. Persona text therefore behaves like a \"few-shot consumable\" that can destroy reliability without touching model weights. All experimental code, raw results, and analysis scripts are available at https://github.com/lexdoudkin/llms-on-drugs.","short_abstract":"Large language models (LLMs) are sensitive to the personas imposed on them at inference time, yet prompt-level \"drug\" interventions have never been benchmarked rigorously. We present the first controlled study of psychoactive framings on GPT-5-mini using ARC-Challenge. Four single-sentence prompts -- LSD, cocaine, alco...","url_abs":"https://arxiv.org/abs/2512.18546","url_pdf":"https://arxiv.org/pdf/2512.18546v1","authors":"[\"Alexander Doudkin\"]","published":"2025-12-21T00:19:02Z","proceeding":"cs.CL","tasks":"[\"cs.CL\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false,"code_links":[{"ID":605743,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2826451,"paper_url":"https://arxiv.org/abs/2512.18546","paper_title":"LLMs on Drugs: Language Models Are Few-Shot Consumers","repo_url":"https://github.com/lexdoudkin/llms-on-drugs","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}