ไปŠๆ—ฅๅทฒๆ›ดๆ–ฐ 305 ๆก่ต„่ฎฏ | ็ดฏ่ฎก 19865 ๆกๅ†…ๅฎน
ๅ…ณไบŽๆˆ‘ไปฌ

๐—”๐—œ ๐—˜๐—ป๐—ด๐—ถ๐—ป๐—ฒ๐—ฒ๐—ฟ๐—ถ๐—ป๐—ด ๐—–๐—ต๐—ฎ๐—ฝ๐˜๐—ฒ๐—ฟ ๐Ÿฏ: ๐—ช๐—ต๐˜† ๐—˜๐˜ƒ๐—ฎ๐—น๐˜‚๐—ฎ๐˜๐—ถ๐—ป๐—ด ๐—”๐—œ ๐—œ๐˜€ ๐—›๐—ฎ๐—ฟ๐—ฑ๐—ฒ๐—ฟ ๐—ง๐—ต๐—ฎ๐—ป ๐—œ๐˜ ๐—Ÿ๐—ผ๐—ผ๐—ธ๐˜€

Yodit Weldegeorgise 2026ๅนด07ๆœˆ07ๆ—ฅ 08:54 4 ๆฌก้˜…่ฏป ๆฅๆบ๏ผšDev.to

One of the biggest takeaways from Chapter 3 of AI Engineering was realizing that building an AI model is only part of the challenge. Figuring out ๐—ต๐—ผ๐˜„ ๐˜๐—ผ ๐—ฒ๐˜ƒ๐—ฎ๐—น๐˜‚๐—ฎ๐˜๐—ฒ ๐—ถ๐˜ ๐—ณ๐—ฎ๐—ถ๐—ฟ๐—น๐˜† ๐—ฎ๐—ป๐—ฑ ๐—ฎ๐—ฐ๐—ฐ๐˜‚๐—ฟ๐—ฎ๐˜๐—ฒ๐—น๐˜† can be just as difficult. With traditional software, it's usually easy to tell whether something works. If a calculation is wrong or a test fails, you know there's a bug. But AI doesn't always work that way. A model can generate multiple reasonable answers to the same question, making it much harder to determine which one is actually better. That made me think: ๐—›๐—ผ๐˜„ ๐—ฑ๐—ผ ๐˜„๐—ฒ ๐—ธ๐—ป๐—ผ๐˜„ ๐—ถ๐—ณ ๐—ฎ๐—ป ๐—”๐—œ ๐—บ๐—ผ๐—ฑ๐—ฒ๐—น ๐—ถ๐˜€ ๐—ฎ๐—ฐ๐˜๐˜‚๐—ฎ๐—น๐—น๐˜† ๐—ถ๐—บ๐—ฝ๐—ฟ๐—ผ๐˜ƒ๐—ถ๐—ป๐—ด? ๐—•๐—ฒ๐—ป๐—ฐ๐—ต๐—บ๐—ฎ๐—ฟ๐—ธ๐˜€ ๐—ก๐—ฒ๐—ฒ๐—ฑ ๐˜๐—ผ ๐—ž๐—ฒ๐—ฒ๐—ฝ ๐—˜๐˜ƒ๐—ผ๐—น๐˜ƒ๐—ถ๐—ป๐—ด Reading this section made me realize how difficult it is for evaluation benchmarks to keep up with the pace of AI development. The chapter explains that GLUE (General Language Understanding Evaluation) was introduced in 2018 to measure how well language models performed on common natural language tasks. But within about a year, models had already become so good at it that researchers introduced SuperGLUE in 2019 as a more difficult benchmark. GLUE evaluates tasks such as: Question answering Sentiment analysis Sentence similarity Text classification The chapter also mentions newer benchmarks like: SuperGLUE MMLU (Massive Multitask Language Understanding) MMLU-Pro Each one was introduced because the previous benchmark was no longer challenging enough. What I found interesting is that a model getting a higher benchmark score doesn't always mean it understands language better. Sometimes it simply means the model has become very good at solving that particular benchmark. ๐—จ๐—ป๐—ฑ๐—ฒ๐—ฟ๐˜€๐˜๐—ฎ๐—ป๐—ฑ๐—ถ๐—ป๐—ด ๐—˜๐—ป๐˜๐—ฟ๐—ผ๐—ฝ๐˜† ๐—ฎ๐—ป๐—ฑ ๐—ฃ๐—ฒ๐—ฟ๐—ฝ๐—น๐—ฒ๐˜…๐—ถ๐˜๐˜† Another section I really enjoyed was the explanation of entropy and perplexity. The chapter explains entropy as a measure of how much information a token carries and how difficult it is to predict the next token in a sequence. Perplexity measures uncertainty. If a model is very uncertain about what comes next, its perplexity will be higher. If

ๆœฌๆ–‡ๅ†…ๅฎนๆฅๆบไบŽไบ’่”็ฝ‘๏ผŒ็‰ˆๆƒๅฝ’ๅŽŸไฝœ่€…ๆ‰€ๆœ‰
ๆŸฅ็œ‹ๅŽŸๆ–‡