Prompt Engineer
LLM fundamentals questions
46 questions on LLM fundamentals for prompt engineers, easiest first.
Easy 24
- What a token iseasy
- What fills the context windoweasy
- Raising the temperatureeasy
- What top-p controlseasy
- Reply cut off mid-sentenceeasy
- Stop sequenceseasy
- The model forgot the last messageeasy
- Questions about last weekeasy
- Message roles in chat APIseasy
- Same text, more tokenseasy
- What you pay foreasy
- Temperature 0 still varieseasy
- Making replies fastereasy
- What top_p controlseasy
- What a stop sequence doeseasy
- Reply ends mid-sentenceeasy
- Estimating tokens from wordseasy
- What presence_penalty doeseasy
- Temperature set to 1.9easy
- Changing both sampling knobseasy
- Counting the r's in a wordeasy
- Brainstorming with temperature 0easy
- What max_tokens limitseasy
- What a 32k context window coverseasy
Medium 7
Hard 15
- Switching to a newer modelhard
- Adding 30 more exampleshard
- Prompt cache hit rate near zerohard
- Reasoning model bill surprisehard
- Fresh facts plus a fixed stylehard
- Long agent run fills contexthard
- Savings from a cached prefixhard
- How much room is left for documentshard
- A more honest confidence scorehard
- Banning a word with logit_biashard
- Caching and time to first tokenhard
- Per-request tool lists and cachinghard
- How many tokens min_p keepshard
- Top-token probability at T = 0.5hard
- Tokens kept at top_p = 0.8hard