Robbie McCorkell
@robbiemccorkell
Followers
389
Following
747
Media
96
Statuses
665
CTO @Leap_Labs. Previously Head of Engineering @salve_health, Tech Lead @RedBadgerTeam
iPhone: 52.021347,-1.737170
Joined July 2008
RT @leedsharkey: Big new review! . 🟦Open Problems in Mechanistic Interpretability🟦. We bring together perspectives from ~30 top researchers….
0
94
0
We built a tool for debugging why your LLM said the thing. For each output token, what input token most influenced it? 🧵below:.
LLM attribution is the most useful prompt engineering tools you've never heard of. It can help you understand and debug model behaviour 🤔
0
0
3
It baffles me that the major cloud providers don't have good solutions for serverless GPU compute yet. I don't want to have to manage VMs or K8s to get access to an A100. Really impressed with @modal_labs so far. Immediate access to H100 level, no approval, pay per second.
1
0
9
Especially as eval techniques become more popular and widespread online. Future models will learn from the latest info on the internet, including the eval techniques, and recognise them as soon as they come up.
0
0
2
"it also highlighted the need for us as an industry to move past artificial tests to more realistic evaluations that can accurately assess models true capabilities and limitations.". Love this. We need tools that give a holistic view of the model.
Fun story from our internal testing on Claude 3 Opus. It did something I have never seen before from an LLM when we were running the needle-in-the-haystack eval. For background, this tests a model’s recall ability by inserting a target sentence (the "needle") into a corpus of
1
0
2
Almost as soon as I tweeted this, @astral_sh once again answered my prayers and released uv for package management. Looks great!
0
0
0
What helps you catch bugs early? What helps your team work together better? What encourages you to better structure your code?.
1
0
0
pdm looks like an interesting (but perhaps early) solution to the mayhem that is pip. Pydantic looks like a nice package for data validation that could help with my zod addiction in JS land. .
2
0
0
Ok tech twitter. I'm doing some Python and want to know the best tools and libraries in the ecosystem. Ruff looks like a great example. Modern, fast, well documented, takes the best of similar projects for other languages etc. What else pushes the language forward to be better?.
1
0
0
I had a great time talking about MLOps Community: LLMs Mini Summit on the @MLOps community Meetup, you can check out the whole presentation here:
home.mlops.community
Deep Dive on LLM Fine-tune In his session, Thomas focuses on understanding the ins and outs of fine-tuning LLMs. We all have a lot of questions during the fine-tuning process. How do you prepare your...
1
2
7
RT @mlopscommunity: Unlock the secrets of maximizing language models' potential at the MLOps Community Mini Summit #3 brought to us by @wei….
0
2
0
RT @aiifyio: Looking forward to exploring our hidden biases (I mean our model's hidden biases 😉) with @JessicaRumbelow & @robbiemccorkell f….
0
1
0
Next on our little Bay Area tour, I'll be showing off the Leap Dashboard at on Saturday.
aiify.io
Aiify.io builds community amongst AI thought leaders, developers, and enthusiasts focussed on discovering and building what's next. Join us to meet and hear from the people and organizations at the...
Don't miss this Saturday's – IRL & Livesteam. RSVP ASAP to secure a spot: — Details: Full Schedule and Speaker Roster are live.
0
1
3
- 🎓 Our research paper on Image Classification Prototypes:.
arxiv.org
We introduce Prototype Generation, a stricter and more robust form of feature visualisation for model-agnostic, data-independent interpretability of image classification models. We demonstrate its...
0
0
0
- 🔔 Join our discord for discussion and news on upcoming releases (including LLM support):.
1
0
0
- 📝 Join the waitlist to run Leap's Prototypes on your own models: .
1
0
0
Thanks for to those that came to AI Tinkerers SF and saw my quick talk on better AI Interpretability. Some quick links for those interested: 🧵.
9/ Leap Labs . Better AI Interpretability. Demystifying models by identifying which features vision models pay attention most to. @robbiemccorkell
2
4
12
RT @atagade19: Can we derive robust insights about learned class features directly from a trained model? We propose Prototype Generation, a….
0
4
0
RT @isnit0: And another two!. Kim Antoniou will share how @auristech is using AI to improve literacy and English language skills for studen….
0
3
0
