meame2010

8 ms·

1.
by meame2010 1mo ago
Thank you. The result that surprised us most while writing it: the ceiling is about 100 reliably triggerable skills, and eviction starts with the ones you have invoked least. So the skill you just installed and never used is the first to lo
2.
New Paper: Attention Is All You Have(twitter.com)
10 points by meame2010 1mo ago|3 comments
3.
by meame2010 7mo ago
which agent did i miss?
4.
by meame2010 7mo ago
thx! contribution is much appreciated!
5.
by meame2010 7mo ago
We built this as a simple open directory: one place to discover, compare, and learn about coding agents. It covers: - coding agents - models - mcp, skills, and protocols - benchmarks - weekly updates This is a community-driven project — the
6.
Show HN: Codingagents.md – The open directory for AI coding agents(codingagents.md)
5 points by meame2010 7mo ago|4 comments
7.
by meame2010 8mo ago
adal supports both terminal and web ui!
8.
by meame2010 8mo ago
thx!
9.
by meame2010 8mo ago
will be out in a week: sign up here: https://sylph.ai/
10.
Show HN: AdaL Web, a local “Claude co-work” [video](youtube.com)
6 points by meame2010 8mo ago|8 comments
11.
CLI agent building on adalflow open-source agent SDK(adal.ml)
2 points by meame2010 1y ago|1 comments
12.
by meame2010 1y ago
Accept waiting list now.
13.
by meame2010 1y ago
links?
14.
Join the team to build the first virtual AI/ML engineer(adal.engineer)
1 points by meame2010 1y ago|3 comments
15.
by meame2010 1y ago
Hiring founding full-stack, AI/ML engineers, and growth lead. If you have an itch for startups or you are an ex-founder, and love the promises of agents and model fine-tuning, you shall be one of us. We are the creators behind AdalFlow
16.
Chat with any GitHub Repo for free [video](youtube.com)
2 points by meame2010 2y ago|1 comments
17.
by meame2010 2y ago
On-going open source project to reach to product-grad product. Built with AdalFlow library: https://github.com/SylphAI-Inc/AdalFlow Will including dataset creation, evaluation, and auto-prompt optimization
18.
by meame2010 2y ago
not few shot, but prompt tuning via text generation via auto-differentiation. https://arxiv.org/abs/2501.16673
19.
A prompt-trained DeepSeek R1 70B can perform better than GPT-o1 using AdalFlow(colab.research.google.com)
4 points by meame2010 2y ago|4 comments
20.
by meame2010 2y ago
Time to move to open-source and smaller reasoning model. Here are the top three learnings from auto-prompt optimizing DeepSeek R1 LLaMA70B for RAG: 1⃣ A trained DeepSeek R1 LLaMA70B(r1 distilled) is even better than GPT-o1 without training.
21.
by meame2010 2y ago
We use gpt4o as the backward model. But I’m excited to try deepseek r1 as it has explicit reasoning available. We are continuously adding more benchmarks to the paper with UTAustin.
22.
by meame2010 2y ago
Yup. The LLM-AutoDiff is just getting started. But it has proven generation-only without explicitly doing few-shot samples can be even more effective and create shorter final prompts
23.
by meame2010 2y ago
Author here. Yea, in this fashion. And it can create the feedback using llm as a backward engine
24.
by meame2010 2y ago
you need a training dataset, and a task pipeline that works. You can refer to this doc: https://adalflow.sylph.ai/use_cases/question_answering.html
25.
Auto-Differentiating Any LLM Workflow: A Farewell to Manual Prompting(arxiv.org)
137 points by meame2010 2y ago|32 comments
26.
by meame2010 2y ago
Implemented in AdalFlow: https://github.com/SylphAI-Inc/AdalFlow
27.
by meame2010 2y ago
Hey Hackers, I'm Li Yin ([GitHub]( https://github.com/liyin2015 )), the author of AdalFlow and a former AI researcher at Meta AI. AdalFlow was inspired by a viral [LinkedIn post]( https://www.linkedin.com/
28.
by meame2010 2y ago
I think you can use the python version to optimize the prompt and use the typescript version to deploy it
29.
by meame2010 2y ago
Thanks for the insightful response. Good point on using 4o-mini to save cost. I'll try it out. I will check more into the soft-prompt tuning. For the current scope, we are focused on in-context learning, ways to improve model reasoning
30.
by meame2010 2y ago
Wow, this is the first time I heard about Ada Language

More ›