A minimal, responsive and SEO-friendly Astro blog theme.https://blog.h.jiaqicai.com/https://blog.h.jiaqicai.com/posts/top-interview-150/https://blog.h.jiaqicai.com/posts/top-interview-150/Worked Python solutions to the LeetCode Top Interview 150, organized by topic with a short approach for each problem.Tue, 02 Jun 2026 14:00:00 GMThttps://blog.h.jiaqicai.com/posts/evolution-of-multimodal-models-alexnet-to-world-models/https://blog.h.jiaqicai.com/posts/evolution-of-multimodal-models-alexnet-to-world-models/A ground-up tour of how neural networks learned to see, then to see-and-read, and finally to imagine. From AlexNet and CNNs, through CLIP and the vision-language models behind GPT-4V, to world models like Dreamer, V-JEPA 2, and LeWorldModel — with architectures, math, and benchmark numbers along the way.Tue, 02 Jun 2026 01:00:00 GMThttps://blog.h.jiaqicai.com/posts/attention-residuals-deep-dive/https://blog.h.jiaqicai.com/posts/attention-residuals-deep-dive/A deep dive into the Kimi team's Attention Residuals (AttnRes) — replacing the fixed-weight residual connection with learned softmax attention over depth. Covers the time–depth duality, Full vs Block AttnRes, the structured-matrix view that unifies prior residual variants, the pipeline-parallel infra that makes it practical, and the scaling-law and 48B-MoE results.Mon, 01 Jun 2026 06:00:00 GMThttps://blog.h.jiaqicai.com/posts/grpo-and-dapo-deep-dive/https://blog.h.jiaqicai.com/posts/grpo-and-dapo-deep-dive/An end-to-end walkthrough of Group Relative Policy Optimization (GRPO) and Decoupled Clip and Dynamic sAmpling Policy Optimization (DAPO) — the two RL algorithms that drive open reasoning models in 2025–2026. Full math, every design choice motivated, and a head-to-head comparison.Thu, 28 May 2026 12:00:00 GMThttps://blog.h.jiaqicai.com/posts/grpo-and-gspo-deep-dive/https://blog.h.jiaqicai.com/posts/grpo-and-gspo-deep-dive/A complete walkthrough of Group Relative Policy Optimization (GRPO) and Group Sequence Policy Optimization (GSPO) — the policy-gradient algorithms behind DeepSeek-R1 and Qwen3. Full math, the failure mode that motivated GSPO, the MoE story, and a side-by-side comparison.Thu, 28 May 2026 12:00:00 GMThttps://blog.h.jiaqicai.com/posts/grpo-and-drgrpo-deep-dive/https://blog.h.jiaqicai.com/posts/grpo-and-drgrpo-deep-dive/An end-to-end derivation of Group Relative Policy Optimization (GRPO) from DeepSeekMath and the Dr.GRPO correction from Liu et al. Covers the full objective, the gradient, the two biases (length and question difficulty), the unbiased fix, and the practical recipe behind R1-Zero–style training.Thu, 28 May 2026 07:00:00 GMThttps://blog.h.jiaqicai.com/posts/training-composer-2/https://blog.h.jiaqicai.com/posts/training-composer-2/A structured walkthrough of Sasha Rush's Training Composer 2 workshop: why Cursor chose Kimi K2.5, how continued pretraining and long-horizon RL fit together, what CursorBench measures, and where Composer is headed.Wed, 27 May 2026 12:00:00 GMThttps://blog.h.jiaqicai.com/posts/leetcode-problems/https://blog.h.jiaqicai.com/posts/leetcode-problems/Leetcode grinding.Tue, 26 May 2026 14:31:00 GMThttps://blog.h.jiaqicai.com/posts/hybrid-attention-and-mla-the-tradeoff/https://blog.h.jiaqicai.com/posts/hybrid-attention-and-mla-the-tradeoff/A side-by-side dive into Xiaomi MiMo's hybrid sliding-window/global attention and DeepSeek's Multi-head Latent Attention. The two answer the same question — how to make attention affordable at long context — with very different bets, and those bets shape everything from training infra to KV cache size.Sat, 23 May 2026 12:00:00 GMThttps://blog.h.jiaqicai.com/posts/kimi-k25-joint-modality-and-agent-swarm/https://blog.h.jiaqicai.com/posts/kimi-k25-joint-modality-and-agent-swarm/A walkthrough of two ideas behind Kimi K2.5: how joint text–vision pre-training and RL make each modality help the other, and how Agent Swarm replaces sequential tool use with a learned parallel orchestrator.Tue, 19 May 2026 12:00:00 GMThttps://blog.h.jiaqicai.com/posts/deepseek-sparse-attention-deep-dive/https://blog.h.jiaqicai.com/posts/deepseek-sparse-attention-deep-dive/A deep dive into DeepSeek's two sparse attention designs — Native Sparse Attention (NSA) and DeepSeek Sparse Attention (DSA) — covering the math, the hardware story, and why DSA in V3.2 looks so different from NSA.Mon, 18 May 2026 12:00:00 GMThttps://blog.h.jiaqicai.com/posts/astro-paper-v6/https://blog.h.jiaqicai.com/posts/astro-paper-v6/AstroPaper v6: a from-scratch rewrite on Astro v6, Tailwind v4, and a new config system.Sun, 17 May 2026 07:15:45 GMThttps://blog.h.jiaqicai.com/posts/how-to-configure-astropaper-theme/https://blog.h.jiaqicai.com/posts/how-to-configure-astropaper-theme/How you can make AstroPaper theme absolutely yours.Sun, 17 May 2026 05:50:08 GMThttps://blog.h.jiaqicai.com/posts/customizing-astropaper-theme-color-schemes/https://blog.h.jiaqicai.com/posts/customizing-astropaper-theme-color-schemes/How you can enable/disable light & dark mode; and customize color schemes of AstroPaper theme.Sun, 17 May 2026 04:57:06 GMThttps://blog.h.jiaqicai.com/posts/adding-new-posts-in-astropaper-theme/https://blog.h.jiaqicai.com/posts/adding-new-posts-in-astropaper-theme/Some rules & recommendations for creating or adding new posts using AstroPaper theme.Sun, 17 May 2026 04:47:08 GMThttps://blog.h.jiaqicai.com/posts/predefined-color-schemes/https://blog.h.jiaqicai.com/posts/predefined-color-schemes/Some of the well-crafted, updated predefined color schemes for AstroPaper.Sat, 16 May 2026 11:29:26 GMThttps://blog.h.jiaqicai.com/posts/dynamic-og-image-generation-in-astropaper-blog-posts/https://blog.h.jiaqicai.com/posts/dynamic-og-image-generation-in-astropaper-blog-posts/New feature in AstroPaper v1.4.0, introducing dynamic OG image generation for blog posts.Mon, 04 May 2026 00:00:00 GMThttps://blog.h.jiaqicai.com/posts/how-to-add-latex-equations-in-blog-posts/https://blog.h.jiaqicai.com/posts/how-to-add-latex-equations-in-blog-posts/Learn how to add LaTeX equations in Astro blog posts using Markdown, KaTeX, and remark/rehype plugins.Sat, 22 Mar 2025 09:25:46 GMThttps://blog.h.jiaqicai.com/posts/how-to-integrate-giscus-comments/https://blog.h.jiaqicai.com/posts/how-to-integrate-giscus-comments/Comment function on a static blog hosted on GitHub Pages with Giscus.Wed, 12 Mar 2025 12:28:53 GMThttps://blog.h.jiaqicai.com/posts/astro-paper-v5/https://blog.h.jiaqicai.com/posts/astro-paper-v5/AstroPaper v5: keep the clean look, updates under the hood.Sat, 08 Mar 2025 08:18:19 GMThttps://blog.h.jiaqicai.com/posts/setting-dates-via-git-hooks/https://blog.h.jiaqicai.com/posts/setting-dates-via-git-hooks/How to use Git Hooks to set your Created and Modified Dates on AstroPaperMon, 08 Jan 2024 18:59:05 GMThttps://blog.h.jiaqicai.com/posts/astro-paper-v4/https://blog.h.jiaqicai.com/posts/astro-paper-v4/AstroPaper v4: ensuring a smoother and more feature-rich blogging experience.Thu, 04 Jan 2024 09:30:41 GMThttps://blog.h.jiaqicai.com/posts/astro-paper-v3/https://blog.h.jiaqicai.com/posts/astro-paper-v3/AstroPaper Version 3: Elevating Your Web Experience with Astro v3 and Seamless View TransitionsMon, 25 Sep 2023 10:25:54 GMThttps://blog.h.jiaqicai.com/posts/how-to-update-dependencies/https://blog.h.jiaqicai.com/posts/how-to-update-dependencies/How to update project dependencies and AstroPaper template.Thu, 20 Jul 2023 15:33:05 GMThttps://blog.h.jiaqicai.com/posts/astro-paper-2/https://blog.h.jiaqicai.com/posts/astro-paper-2/AstroPaper with the enhancements of Astro v2. Type-safe markdown contents, bug fixes and better dev experience etc.Mon, 30 Jan 2023 15:57:52 GMThttps://blog.h.jiaqicai.com/posts/examples/tailwind-typography/https://blog.h.jiaqicai.com/posts/examples/tailwind-typography/EXAMPLE POST: About Tailwind Typography Plugin and how you can use it effectively.Tue, 05 Jul 2022 02:05:51 GMThttps://blog.h.jiaqicai.com/posts/examples/how-do-i-develop-my-terminal-portfolio-website-with-react/https://blog.h.jiaqicai.com/posts/examples/how-do-i-develop-my-terminal-portfolio-website-with-react/EXAMPLE POST: Developing a terminal-like website using ReactJS, TypeScript and Styled-Components. Includes features like autocomplete, multiple themes, command hints etc.Thu, 09 Jun 2022 03:42:51 GMThttps://blog.h.jiaqicai.com/posts/examples/how-do-i-develop-my-portfolio-and-blog/https://blog.h.jiaqicai.com/posts/examples/how-do-i-develop-my-portfolio-and-blog/EXAMPLE POST: My experience about developing my first portfolio website and a blog using NextJS and a headless CMS.Fri, 25 Mar 2022 16:55:12 GMT