Initially I aimed to test with at least 10 formulas for each model for SAT/UNSAT, but it turned out to be more expensive than I expected, so I tested ~5 formulas for each case/model. First, I used the openrouter API to automate the process, but I experienced response stops in the middle due to long reasoning process, so I reverted to using the chat interface (I don't if this was a problem from the model provider or if it's an openrouter issue). For this reason I don't have standard outputs for each testing, but I linked to the output for each case I mentioned in results.
Publication date: 28 February 2026。业内人士推荐谷歌浏览器【最新下载地址】作为进阶阅读
В России ответили на имитирующие высадку на Украине учения НАТО18:04，推荐阅读服务器推荐获取更多信息
ALiBi enables extreme compression: the 36-param leader uses ALiBi with slope log(10) for base-10 positional weighting, achieving 100% accuracy with a 2-layer decoder (d=5) in float64，推荐阅读夫子获取更多信息