Hi everyone! 👋 I'm kicking-off a journey to learn testing AI/ML systems & share the whole adventure. Follow along! Also check my activity on LinkedIn: https://t.co/A1LLWigMjD
#AI#ML#Testing#AITestingQuest
Next up, I'm exploring the wider world of machine learning workflows 🤖⚙️ and how to automate and test each step. Stay tuned for more insights!
#AI#LLM#MachineLearning#Testing#AIWorkflow
After a short break, I'm excited to share my latest progress with testing LLMs. 🎓Check out my learning journey, live sessions, and code here: ▶️ LLM Testing Playlist: https://t.co/xFaNaKFmAf ▶️ Live Recordings: https://t.co/hoOMs7CMrb ▶️ Repo & Resources: https://t.co/wJPjsJ39VA
💥KA-POW💥
🤖Prompt Engineering🤖
Designing and optimizing inputs (prompts) to elicit the best responses from language models. It's like crafting the perfect question to get a detailed answer. #AI#LLM#KAPOW
💥KA-POW💥
🤖BERT🤖
Bidirectional Encoder Representations from Transformers, a model that understands context in all directions. It's like reading a sentence forward and backward for deeper meaning. #AI#LLM#KAPOW
💥KA-POW💥
🤖Self-Supervised Learning🤖
Learning from unlabeled data by generating labels from the data itself. It's like learning to solve puzzles by understanding patterns. #AI#LLM#KAPOW
Since Friday's live, I've been sprinting towards the LLM testing knowledge with my dorky hat 🏃♂️💨. I've hit a few bumps along the way and even faceplanted 😅, but I'm still determined, not a chance to give up.
💥KA-POW💥
🤖Distillation🤖
Compressing a large model into a smaller one while retaining performance. Think of it as creating a pocket-sized version of an encyclopedia. #AI#LLM#KAPOW
Had an awesome Friday Live session!🎥We "dived deep into the ever-changing world of LLMs, exploring their potential and limitations". 🤖 Started scripting an LLM judge evaluation—avoiding (mostly) other LLMs for fairness!😅
Hi everyone! 👋
Take a different approach to the same question we already discussed about. Now with a bit more understanding and experience👉https://t.co/oojVXntTLZ
The place and time will be similar as last time:
Youtube, AITestingQuest channel at UTC 18:00 / EEST 21:00.
💥KA-POW💥
🤖Positional Encoding🤖
Adding information about the order of words in a sequence to help the model understand context. It's like numbering the pages of a book. #AI#LLM#KAPOW
🚀Dive into AI testing with my latest video! I tested LLMs using 7 metrics and a factual correctness check, and the results are eye-opening.😱💥
📊phi3: 6/10 pass➡️4/7 metrics📊stablelm2: 6/10 pass➡️5/7 metrics📊tinyllama: 4/10 pass➡️5/7 metrics
Link in the comment! ⬇️⬇️⬇️⬇️⬇️
💥KA-POW💥
🤖Masked Language Modeling🤖
A training method where some words in a sentence are hidden, and the model learns to predict them. It's like filling in the blanks of a story. #AI#LLM#KAPOW
Evidently AI is a versatile tool for LLM & ML evaluation, mixing NLTK, DeepEval, and more. Check their docs: https://t.co/pt5BBZR9hr
Next session: 30.08.2024 @ 18:00 UTC - Testing an LLM | LLM evaluating LLMs: Can We Trust These AI Judges?
#LLM#AI#Testing#DeepEval#EvidentlyAI
🚨 We just wrapped up the tool exploration subseries of "Testing an LLM"! 🎉 Last session was all about Evidently AI, and it was a wild ride! 🎢
Missed it? I’ll have a shorter version ready soon. But if you’re into the live drama, remember: Fridays are for #AITestingQuest! 😉
Don't have any plans for today? Even if you do, just cancel it and join our next live session 👉https://t.co/NBqkwdEi4U
Book in your calendar: UTC 18:00 / EEST 21:00.
Evidently AI is in the focus this time, let's see how it works!🖖
#AITestinQuest#YoutubeLive#LLM#EvidentlyAI
💥KA-POW💥
🤖Beam Search🤖
A heuristic search algorithm that explores multiple possible sentences simultaneously to find the best one. It's like navigating multiple paths to find the quickest route home. #AI#LLM#KAPOW