https://jinkunchen.com/blog Jinkun's Blog en https://jinkunchen.com/blog/when-an-ai-agent-says-i-who-are-we-trusting https://jinkunchen.com/blog/when-an-ai-agent-says-i-who-are-we-trusting Sun, 26 Apr 2026 00:00:00 GMT Personified AI agents can make users lose track of who is deciding, who is authorizing, and who is responsible. https://jinkunchen.com/blog/context-order-and-reasoning-drift-measuring-order-sensitivity-from-token-probabilities https://jinkunchen.com/blog/context-order-and-reasoning-drift-measuring-order-sensitivity-from-token-probabilities Sat, 07 Feb 2026 00:00:00 GMT How the same evidence can yield different answers when you only change the order. https://jinkunchen.com/blog/measuring-reasoning-drift-a-token-level-instability-signal https://jinkunchen.com/blog/measuring-reasoning-drift-a-token-level-instability-signal Fri, 06 Feb 2026 00:00:00 GMT How to detect process-level breakdowns from token probabilities before the final answer fails. https://jinkunchen.com/blog/when-ai-reasoning-starts-to-drift https://jinkunchen.com/blog/when-ai-reasoning-starts-to-drift Mon, 05 Jan 2026 00:00:00 GMT People often describe AI mistakes as sudden failures. One moment the answer seems reasonable and confident, and the next it is clearly wrong. This makes erro… https://jinkunchen.com/blog/the-effect-of-chunk-retrieval-sequence-in-rag-on-multi-step-inference-performance-of-large-language-models https://jinkunchen.com/blog/the-effect-of-chunk-retrieval-sequence-in-rag-on-multi-step-inference-performance-of-large-language-models Thu, 19 Jun 2025 00:00:00 GMT Why the order of retrieved information can quietly change how AI reasons step by step https://jinkunchen.com/blog/key-challenges-in-current-llm-memory-systems https://jinkunchen.com/blog/key-challenges-in-current-llm-memory-systems Tue, 17 Jun 2025 00:00:00 GMT Why how AI remembers information can quietly shape its reasoning https://jinkunchen.com/blog/the-surprising-impact-of-memory-order-on-llm-responses https://jinkunchen.com/blog/the-surprising-impact-of-memory-order-on-llm-responses Mon, 09 Jun 2025 00:00:00 GMT Why the sequence of information can quietly change how AI reasons https://jinkunchen.com/blog/do-language-model-embeddings-form-an-approximate-abelian-group https://jinkunchen.com/blog/do-language-model-embeddings-form-an-approximate-abelian-group Tue, 03 Jun 2025 00:00:00 GMT This post explores a theoretical question about how language models represent meaning internally. The main text uses mathematical concepts to examine whether…