6/ Teaching LLM Agents to Self-Improve – claims it is possible to iteratively fine-tune LLMs with the ability to improve their own response over multiple turns with additional environment feedback; the LLM learns to detect and correct its previous mistakes in subsequent
PROMPT ENGINEERING
-
RAG vs Long-Context LLMs: Self-Route Strategy
By
–
3/ RAG vs. Long-Context LLMs – compares RAG and long-context LLMs and finds that long-context LLMs outperform RAG on average performance while RAG is significantly less expensive; proposes Self-Route, leveraging self-reflection to route queries to RAG or LC; reports that
-
Good Instruction Following Test with Temperature 1.0 Sampling
By
–
Actually a pretty good instruction following test. Ideally I’d run 10 samples/model with temperature 1.0 and all other bells and whistles (topp/k) off.
-
Interactive LLM Decoding Strategies and Sampling Effects
By
–
Interactive LLM decoding strategies LLM sampling is a cool app that shows the effect of temperature, top K sampling, top P sampling, etc. on the probability distribution outputted by an LLM. LLM sampling: https://
artefact2.github.io/llm-sampling/i
ndex.xhtml
… -
OpenAI’s SearchGPT vs Perplexity AI
By
–
OpenAI introduced SearchGPT.
— God of Prompt (@godofprompt) 26 juillet 2024
It's like Perplexity AI but more powerful.
Here's everything you need to know a minute: pic.twitter.com/LNGs0n8CwOOpenAI introduced SearchGPT. It's like Perplexity AI but more powerful. Here's everything you need to know a minute:
-
ChatGPT Classic gives 3 different R counts for ‘strrrraaawwberrrrrrry’
By
–
I asked ChatGPT Classic 3 times and it said 9, 10, and 11. Prompt: How many R’s are in “strrrraaawwberrrrrrry”?
-
Counting R’s in ‘strawberry’ with chain-of-thought as LLM psychology
By
–
There’s no reason to do this except as an exercise in LLM psychology, but here’s how you could count how many R’s are in “strawberry” using chain-of-thought. (Normally, letter- or count-based tasks like this would be handled outside the LLM, by having it call an API or REPL.)
-
LLM Tokenization Explained Through Emoji Translation
By
–
To help explain the weirdness of LLM Tokenization I thought it could be amusing to translate every token to a unique emoji. This is a lot closer to truth – each token is basically its own little hieroglyph and the LLM has to learn (from scratch) what it all means based on
-
Chat with Your Data: Summarizing PDF Files in Apps
By
–
Step 4: Chat with your data Finally, once the data has been ingested into the app, chat with your data. Here we are asking it to summarize the pdf file.