https://www.apolloresearch.ai/science/stress-testing-deliberative-alignment-for-anti-scheming-training/ 2026-04-21T11:12:18+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/09/Group-47-1.png https://www.apolloresearch.ai/wp-content/uploads/2025/09/figure_one_square.webp https://www.apolloresearch.ai/wp-content/uploads/2025/09/image4.png https://www.apolloresearch.ai/wp-content/uploads/2025/09/image7.png https://www.apolloresearch.ai/wp-content/uploads/2025/09/image9.png https://www.apolloresearch.ai/wp-content/uploads/2025/09/image2.png https://www.apolloresearch.ai/wp-content/uploads/2025/09/image3.png https://www.apolloresearch.ai/wp-content/uploads/2025/09/image1.png https://www.apolloresearch.ai/wp-content/uploads/2025/09/image6.png https://www.apolloresearch.ai/wp-content/uploads/2025/09/image8-1.png https://www.apolloresearch.ai/science/more-capable-models-are-better-at-in-context-scheming/ 2026-04-21T11:26:07+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/10/Evals_specific.png https://www.apolloresearch.ai/science/research-note-our-scheming-precursor-evals-had-limited-predictive-power-for-our-in-context-scheming-evals/ 2026-04-21T11:29:37+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/10/Evals_specific.png https://www.apolloresearch.ai/science/claude-sonnet-37-often-knows-when-its-in-alignment-evaluations/ 2026-04-21T11:32:57+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/10/Evals_specific.png https://www.apolloresearch.ai/science/forecasting-frontier-language-model-agent-capabilities/ 2026-04-21T11:36:56+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/10/Evals_specific.png https://www.apolloresearch.ai/wp-content/uploads/2025/08/forecasting_1.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/forecasting_2.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/forecasting_5.webp https://www.apolloresearch.ai/science/interpretability-in-parameter-space-minimizing-mechanistic-description-length-with-attribution-based-parameter-decomposition/ 2026-04-21T11:40:19+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/02/Group-5-scaled.png https://www.apolloresearch.ai/science/detecting-strategic-deception-using-linear-probes/ 2026-04-21T11:42:13+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/02/Group-11-2-e1760437712944.png https://www.apolloresearch.ai/science/demo-example-scheming-reasoning-evaluations/ 2026-04-21T12:11:46+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/10/Evals_specific.png https://www.apolloresearch.ai/science/the-evals-gap/ 2026-04-21T12:20:26+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/10/Evals_broad.png https://www.apolloresearch.ai/wp-content/uploads/2025/08/evals_gap.webp https://www.apolloresearch.ai/science/frontier-models-are-capable-of-incontext-scheming/ 2026-04-21T12:20:47+00:00 https://www.apolloresearch.ai/wp-content/uploads/2024/12/Group-11-scaled-e1760437810863.png https://www.apolloresearch.ai/wp-content/uploads/2025/08/01_opus_scheming_headline_figure.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/02_scheming_results_table.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/03_scheming_follow_up_deception_rates.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/04_models_understand_scheming.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/05_no_goal_nudging_description.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/06_instrumental_alignment_faking.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/07_sandbagging_table.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/08_sandbagging_helpful_goal.webp https://www.apolloresearch.ai/science/towards-safety-cases-for-ai-scheming/ 2026-04-21T12:24:28+00:00 https://www.apolloresearch.ai/wp-content/uploads/2024/10/Group-6-2-scaled.png https://www.apolloresearch.ai/wp-content/uploads/2025/08/Temp402x28529.webp https://www.apolloresearch.ai/science/chain-of-thought-monitorability-a-new-and-fragile-opportunity-for-ai-safety/ 2026-04-21T12:27:22+00:00 https://www.apolloresearch.ai/science/an-opinionated-evals-reading-list/ 2026-04-21T12:29:49+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/10/Evals_broad.png https://www.apolloresearch.ai/science/identifying-functionally-important-features-with-end-to-end-sparse-dictionary-learning/ 2026-04-21T12:32:06+00:00 https://www.apolloresearch.ai/wp-content/uploads/2024/05/Group-8-scaled.png https://www.apolloresearch.ai/wp-content/uploads/2025/08/Screenshot2024-05-24at12.28.21.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/Screenshot2024-05-30at11.36.04.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/Screenshot2024-05-17at15.39.21.webp https://www.apolloresearch.ai/science/the-local-interaction-basis-identifying-computationally-relevant-and-sparsely-interacting-features-in-neural-networks/ 2026-04-21T12:36:36+00:00 https://www.apolloresearch.ai/wp-content/uploads/2024/05/Group-14-2-scaled-e1760437876515.png https://www.apolloresearch.ai/wp-content/uploads/2025/08/lib_post_fig1.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/lib_post_fig2.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/lib_post_fig3.webp https://www.apolloresearch.ai/science/black-box-access-is-insufficient-for-rigorous-ai-audits/ 2026-04-21T12:41:09+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/10/Evals_specific.png https://www.apolloresearch.ai/science/we-need-a-science-of-evals/ 2026-04-21T12:43:21+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/10/Evals_broad.png https://www.apolloresearch.ai/wp-content/uploads/2025/01/Frame17.webp https://www.apolloresearch.ai/science/a-starter-guide-for-evals/ 2026-04-21T12:44:51+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/10/Evals_broad.png https://www.apolloresearch.ai/science/large-language-models-can-strategically-deceive-their-users-when-put-under-pressure/ 2026-04-21T15:50:04+00:00 https://www.apolloresearch.ai/wp-content/uploads/2023/11/Large-Language-Models-can-Strategically-Deceive-their-Users-when-Put-Under-Pressure.png https://www.apolloresearch.ai/science/a-causal-framework-for-ai-regulation-and-auditing/ 2026-04-21T15:52:17+00:00 https://www.apolloresearch.ai/wp-content/uploads/2023/11/Group-10-3-1-scaled-e1760437901484.png https://www.apolloresearch.ai/wp-content/uploads/2025/08/reversed_causal_chain_online.webp https://www.apolloresearch.ai/science/our-research-on-strategic-deception-presented-at-the-uks-ai-safety-summit/ 2026-04-21T15:57:01+00:00 https://www.apolloresearch.ai/wp-content/uploads/2023/11/Group-9-1-scaled.png https://www.apolloresearch.ai/science/understanding-strategic-deception-and-deceptive-alignment/ 2026-04-21T16:00:43+00:00 https://www.apolloresearch.ai/wp-content/uploads/2025/10/Evals_broad.png https://www.apolloresearch.ai/wp-content/uploads/2025/08/Frame4.webp https://www.apolloresearch.ai/wp-content/uploads/2025/08/Frame24.webp https://www.apolloresearch.ai/science/metagaming-matters-for-training-evaluation-and-oversight/ 2026-04-22T11:20:20+00:00 https://www.apolloresearch.ai/wp-content/uploads/2026/03/Metagaming.png https://www.apolloresearch.ai/science/science-of-scheming/ 2026-04-23T10:04:37+00:00 https://www.apolloresearch.ai/wp-content/uploads/2026/01/Science_of_Scheming-1.svg https://www.apolloresearch.ai/wp-content/uploads/2026/01/image.png https://www.apolloresearch.ai/wp-content/uploads/2026/01/image-1.png https://www.apolloresearch.ai/wp-content/uploads/2026/01/image-2.png https://www.apolloresearch.ai/wp-content/uploads/2026/01/image-3.png https://www.apolloresearch.ai/wp-content/uploads/2026/01/image-6.png https://www.apolloresearch.ai/wp-content/uploads/2026/01/image-5.png