{"id":"6a75a48ea05e","page":"/c/6a75a48ea05e/","topic":"Reduction in AI scheming rates after deliberative alignment","publisher":"OpenAI","official":true,"platform":"X","publishedAt":"2025-09-17T17:09:36.000Z","model":"GPT mini models","form":"benchmark-bars","metrics":{"views":42468,"likes":289,"reposts":17,"quotes":5,"replies":2,"saves":32},"readAt":"2026-10-12T21:30:08.589Z","chartsInPost":1,"score":50,"scoreWeights":{"views":0.4,"likes":0.3,"eng":0.15,"save":0.15},"quadrant":"hit","ranks":{"score":{"industry":{"n":618,"of":1162},"company":{"n":50,"of":74}},"views":{"industry":{"n":568,"of":1162},"company":{"n":58,"of":74}},"likes":{"industry":{"n":580,"of":1162},"company":{"n":53,"of":74}},"eng":{"industry":{"n":577,"of":1162},"company":{"n":23,"of":74}},"save":{"industry":{"n":589,"of":1162},"company":{"n":15,"of":74}}},"percentiles":{"industry":{"views":0.5116179001721171,"likes":0.5,"eng":0.5038726333907056,"save":0.49354561101549055},"company":{"views":0.22297297297297297,"likes":0.2905405405405405,"eng":0.6959459459459459,"save":0.8040540540540541}},"peers":{"industry":1162,"company":74,"companyName":"OpenAI"},"medians":{"industry":{"views":39945,"likes":289,"eng":0.007331788144862199,"save":0.0007642791742968587},"company":{"views":161517.5,"likes":714.5,"eng":0.005359315036129835,"save":0.00048697197603225794},"form":{"n":215,"views":44446,"eng":0.007500781331388686}},"ratios":{"vsCompanyViews":0.262931261318433,"vsCompanyLikes":0.40447865640307906,"viewsPerFollower":0.007816590984931326,"viewsPerDay":108.84181269812663},"followers":5433059,"reads":[{"at":"2026-10-12T15:48:24.057Z","views":42468,"likes":289,"reposts":17,"quotes":5,"replies":2,"saves":32,"source":"x-search-cache"},{"at":"2026-10-12T15:49:16.604Z","views":42468,"likes":289,"reposts":17,"quotes":5,"replies":2,"saves":32,"source":"x-search-cache"},{"at":"2026-10-12T20:03:09.416Z","views":42468,"likes":289,"reposts":17,"quotes":5,"replies":2,"saves":32,"source":"x-search-cache"},{"at":"2026-10-12T20:04:29.853Z","views":42468,"likes":289,"reposts":17,"quotes":5,"replies":2,"saves":32,"source":"x-search-cache"},{"at":"2026-10-12T20:46:03.444Z","views":42468,"likes":289,"reposts":17,"quotes":5,"replies":2,"saves":32,"source":"x-search-cache"}],"chartData":"/api/charts/6a75a48ea05e.json","copy":{"en":{"why":[{"factor":"views-per-follower","text":"Its 0.008 views per follower sits at the 3rd percentile across posts with known followers, in the bottom decile."},{"factor":"reach-vs-company","text":"At 42.5K views it runs 0.26x OpenAI's 162K median, ranking #58 of 74 OpenAI posts."},{"factor":"rank","text":"Its score of 50 lands at #618 of 1,162 AI chart posts and #50 of 74 OpenAI posts."}],"verdict":"The post sits in the Reached and resonated quadrant at #618 of 1,162 AI chart posts.","news":{"headline":"OpenAI posts anti-scheming benchmark chart for o4-mini and o3","dek":"A benchmark bar chart on September 17, 2025 shows scheming rates for o4-mini and o3 before and after anti-scheming training.","what":"OpenAI published a benchmark bar chart on September 17, 2025 on reduction in AI scheming rates after deliberative alignment training. It was the account's only post that day.","numbers":"It drew 42.5K views, ranking #568 of 1,162 AI chart posts and #58 of 74 OpenAI posts. Its score of 50 ranks #618 of 1,162 AI chart posts.","lesson":"A benchmark bar chart with 4 models shown drew 42.5K views against a 39.9K median among AI chart posts. Its save rate of 0.075% and engagement rate of 0.74% are the measured engagement against saves for this post.","implication":"At #50 of 74 OpenAI posts by score, this chart sits below the account's own median. The 0.26x gap to OpenAI's 162K median views is the measured distance for this publisher's chart posts."}},"zh":{"why":[{"factor":"views-per-follower","text":"浏览量4.2万，粉丝543万，单粉浏览0.008，低于中位数0.20，排第3百分位。"},{"factor":"reach-vs-company","text":"浏览量4.2万，仅为公司中位数16.2万的0.26倍，在74条帖子中排第58。"},{"factor":"rank","text":"综合得分50，在1162条AI图表帖中排第618，在74条OpenAI帖子中排第50。"}],"verdict":"传得远，也打动人，但综合排名在1162条中排第618。","news":{"headline":"OpenAI发布对齐降欺骗率图表","dek":"浏览量4.2万，互动率0.74%，低于公司中位数。","what":"OpenAI于2025年9月17日发布1张基准柱状图，展示4个模型的对齐训练前后欺骗率。当日无其他帖子。","numbers":"浏览量4.2万，在1162条AI图表帖中排第568，低于中位数4万。点赞289，互动率0.74%，收藏率0.075%。","lesson":"浏览量低于公司中位数16.2万的0.26倍，在74条帖子中排第58。单粉浏览0.008，排第3百分位，属后10%。","implication":"该帖在OpenAI自身帖子中排名靠后，但互动率0.74%不低于AI图表帖中位数，属于传得远也打动人象限。"}},"lane":"en:amd/DeepSeek-V4.1-Flash zh:amd/DeepSeek-V4-Flash","checkedAt":"2026-10-13T20:40:06.483Z"}}