一项名为“递归预测”的新提案旨在从主要为短期奖励进行优化的AI模型中引出准确的长期预测。该方法不是直接询问遥远的结果，而是要求AI预测其自身的未来预测，并利用这些中间预测来提供奖励。这个过程创建了一个可验证的短期预测链，最终导致更可靠的长期预测。该方法由Arun和Alex在一篇博文中详细介绍，其他几位研究人员也做出了贡献。 AI
影响 该方法可以提高AI在长期规划和决策中的预测可靠性。
排序理由 这是一篇提出AI对齐新方法的论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
<p><span>We’d like to use powerful AIs to answer questions that may take a long time to resolve. But if a model only cares about performing well in ways that are verifiable shortly after answering (e.g., a myopic </span><a href="https://www.lesswrong.com/posts/bhtYqD4FdK6AqhFDF/f…
<p><span>We’d like to use powerful AIs to answer questions that may take a long time to resolve. But if a model only cares about performing well in ways that are verifiable shortly after answering (e.g., a myopic </span><a href="https://www.lesswrong.com/posts/bhtYqD4FdK6AqhFDF/f…