Long context, together with almost perfect retrieval and memory, would do that, plus sprinkle in a good level of understanding.
@skirano
-
MLX Axolotl Google Vertex AI frameworks comparison
By
–
MLX, Axolotl, Google vertex, depends on the cases
-
Converting JSONL files to TXT format for upload
By
–
I had them in JSONL, but you can't upload that type of file, so I converted the files to TXT.
-
Data Structure Strategy for Model Fine-tuning
By
–
That's just how I structure my data when I do a regular fine-tune; it's just random samples across two files for training and validation.
-
One-Shot Prompts for Fine-Tuning Models Without Extra Instructions
By
–
I could say to avoid that, I do use a lot of emojis though, but for a one-shot prompt to fine-tune with no extra instructions is super impressive.
-
Fine-tuning Hundreds of LLMs Achieves Best Tweet Results
By
–
For context, I have fine-tuned hundreds of large language models; these are the best results I have seen on my tweets.
-
Claude 3’s Retrieval Ability Mimics Fine-tuning with Structured Data
By
–
Claude 3's retrieval ability over long content is so good that if you provide structured data, it essentially acts like a fine-tune! I gave it 800 of my tweets and now it just speaks like me. It's hard to convey my level of shock… It even tagged @Replit and @everartai
! -
Be Careful Pushing Claude: Account Blocking Risks
By
–
Unfortunately, you need to be careful when pushing Claude, as they may block your account, annoying.
-
Testing One-Shot Prompting with Page Source Knowledge
By
–
I agree! I wanted to test how far I could go with one-shot prompting and knowing the page source.
-
Cursor API Integration and Claude Function Calling Capabilities
By
–
I just use the chat, and yes, if Cursor could add the API, it would be huge. I do believe Claude doesn't support function calling right now, so the implementation of some features they have may be less trivial.