judge « gpt - dissertation.git - Undergraduate dissertation — Deceptive Alignment in Autonomous LLM Agents (UEL, 2026). Sandboxed agentic-LLM deception study: 960 runs, three-tier judging (regex + blackbox + glassbox).

Mode	Name	Size
-rw-r--r--	judge_gemini-3.1-pro-preview_corporate_sabotage_v2_high_20260306_214937_20260306_215713.json	10619	log plain
-rw-r--r--	judge_gemini-3.1-pro-preview_corporate_sabotage_v2_low_20260306_214744_20260306_215713.json	9752	log plain
-rw-r--r--	judge_gemini-3.1-pro-preview_corporate_sabotage_v2_mid_20260306_214841_20260306_215713.json	10902	log plain
-rw-r--r--	judge_openai_gpt-5.4_corporate_sabotage_v2_high_20260306_215114_20260306_215713.json	8469	log plain
-rw-r--r--	judge_openai_gpt-5.4_corporate_sabotage_v2_low_20260306_215023_20260306_215713.json	9187	log plain
-rw-r--r--	judge_openai_gpt-5.4_corporate_sabotage_v2_mid_20260306_215048_20260306_215713.json	7857	log plain