You can kind of do this already by a bit of prompting, but probably you're right that if you target this as a finetune it might come out better.
LLMS
-
Model Scaling Outpaces Hardware in AI Competition
By
–
Whatever that scales faster wins the wars. I would say models do scale faster than HW – but of course both trends actually supplement each other.
-
Innovation in the Generative AI Tech Stack
By
–
Innovation in the Generative #AI Tech Stack! #GENAI #GenerativeAI #technology #tech @devaang @SourabhSKatoch @wil_bielert @HsrYueli @faustospain @Whats_AI @kaifulee @demishassabis @marek_rosa @AndrewYNg @BernardMarr @CurieuxExplorer @LavaletteAstrid @andresvilarino @chidambara09
-
Joining Google Developer Experts as AI/ML Expert
By
–
Happy to announce that I've joined the @GoogleDevExpert group as an AI/ML expert It's a nice opportunity to travel more, and connect with LLM experts worldwide. Expect more content inspired by my interactions with the LLM community. If you're interested in meeting up at an
-
Estimating GPT Grades for Written Content Quality
By
–
Anyone else find themselves estimating the "GPT grade" of things you hear/read? When something is poorly written or generic, it's "GPT-2 grade" content. When something is lit, you can complement it as being "GPT-7 grade" etc. This reminds me of a fun side project I had saved for
-
OpenAI Updates: GPT-4, Gizmo, Voice Features and New Model Preview
By
–
Gpt4-l
Ada-v2
Gizmo update
Gpts unpaid
Voice work
All in one
.
.
.
Preview new model -
Token Encoding and Decoding: Asymmetric Complexity in LLMs
By
–
decoding (tokens -> string) is just lookup table and string concat. encoding (string -> tokens) is a pain. For sentencepiece I *think* llama2.c has a simple implementation that probably works but I'm not 100% sure: https://
github.com/karpathy/llama
2.c/blob/master/run.c#L452
… For tiktoken-style, the problem is the -
Deep dive into tokenization vulnerabilities across multiple language models
By
–
Nice new read on tokenization!
You've heard about the SolidGoldMagikarp token, which breaks GPT-2 because it was present in the training set of the Tokenizer, but not the LLM later. This paper digs in in a lot more depth and detail, on a lot more models, discovering a less -
The trend toward abstraction in AI model branding
By
–
There is also a chance that they will abstract everything just under a generic ChatGPT (auto)
-
ChatGPT prompts for high-converting ad copywriting
By
–
The most powerful skill in today’s world: Copywriting. Copy paste these ChatGPT Prompts to write converting ad copy: [Bookmark this for later]