From e53739367ae6053b3ac05ce19ec0cae389521521 Mon Sep 17 00:00:00 2001 From: Shayan Rais Date: Thu, 7 May 2026 11:45:00 +0500 Subject: [PATCH] =?UTF-8?q?add=20llm-basic.svg=20=E2=80=94=20animated=20au?= =?UTF-8?q?toregressive=20generation=20diagram?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Three-panel SVG (input context, LLM black box, predicted next token) with 7-iteration loop generating "The capital of Japan is Tokyo." from the prompt "What is the capital of Japan?". Includes purple feedback loop showing each predicted token appended back into the input. Co-Authored-By: Claude --- presentation/assets/llm/llm-basic.svg | 123 ++++++++++++++++++++++++++ 1 file changed, 123 insertions(+) create mode 100644 presentation/assets/llm/llm-basic.svg diff --git a/presentation/assets/llm/llm-basic.svg b/presentation/assets/llm/llm-basic.svg new file mode 100644 index 0000000..1e77767 --- /dev/null +++ b/presentation/assets/llm/llm-basic.svg @@ -0,0 +1,123 @@ + + + + + + + + + + + + + + + How an LLM generates text (autoregressive) + + + Each predicted token is appended to the input, then fed back into the LLM + + + ITERATION 1 / 7 + ITERATION 2 / 7 + ITERATION 3 / 7 + ITERATION 4 / 7 + ITERATION 5 / 7 + ITERATION 6 / 7 + ITERATION 7 / 7 + + + INPUT (CONTEXT) + + + Original prompt + What + is + the + capital + of + Japan + ? + + Generated tokens (autoregressive) + The + capital + of + Japan + is + Tokyo + . + + + + + + + + LLM + (black box) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + PREDICTED NEXT TOKEN + + argmax P(next token | input) + The + capital + of + Japan + is + Tokyo + . + + + + + + + predicted token appended to input → next iteration + +