Explore tweets tagged as #LLMHack
Your AI Assistant Might Be Leaking Secrets. Prompt injection now tops the list of 2025’s AI security threats. Bad actors can trick AI into revealing data or performing malicious tasks. #PromptInjection #AISecurity #LLMHack #AIThreat #OWASP #CyberRisk
0
0
0
As a part of the #llmhack, me and Sean Warren developed MolVerse, a language interface to 3dmol.js: a web app for all chemists and biologists that want to visualise their (bio)molecules without needing to code. Here we load the structure of benzene:
5
32
176
And here is our text-to-CAD demo that got us 🥈second place at the LLMHack by EF with Mistral and HF.It's an agent that writes code for the CAD model generation.I build the agent for this over the hackathon — and also went to a nice physics TD, a bio exam, and got a 92 sleep
Two days ago my friend told me that since I'm writing in 🦀 @rustlang (my physics simulation was faster than his Fortran one) I need programming socks (long anime socks or sth). And now I was participating in this @join_ef + @MistralAI + @huggingface #llmhackparis, we built
5
7
24
It's a warp!. - 53 authors.- 14 projects. after 1.5 days of hackathon. With this paper, we have a full bouquet of examples of how LLMs can impact chemistry, materials science, and beyond. #llmhack #llm #chemtwitter
3
56
243
Kevin Jablonka described his latest work on LLMs applied to chemical discovery. He is building software to make using these models ridiculously simple. @kmjablonka #llmhack @EPFL_en . Recording available later
1
4
33
Chris Dallago from @nvidia describing some of their latest tools to help researchers apply LLMs to chemical and biological applications and MegaMolBART sampling. @sacdallago #OpenScience #llmhack @anthonycosta
1
7
14
1/n) Include any chemical feature into an LLM for chemistry 🤯How? introducing our hackathon project "geometric-geniuses" - get the best of both worlds! numerical feature vectors inside of LLMs.#LLMhack
1
3
28
Who’s attending the LLM Hack by @join_ef @MistralAI @huggingface in Paris?. #llmhack #ai #hackathon.
0
0
1
One hack I use a lot when I am stuck is to take a screenshot of that problem and send it to whatever LLM I am using; I use Gemini for this. It pinpoints the error and points me in the specific direction needed. A lifesaver, never failed me. #automation .#llmhack.
1
0
4
This is @jablonkagroup's submission. Very happy to have worked with @Iam_Nawaf_, Mara and Abdel. Also, thanks a lot to the #LLMHack organizers, the @FAIRmat_NFDI for having us and all the people who attend in Berlin. #LLMs4Science.
0
0
5
My first foray into fine-tuning LLMs to convert textual descriptions of organic procedures into structured JSON data suitable for downstream ML tasks. Was exciting to work with @QaiAlex and the rest of the team at #llmhack and fine-tune the LLM!.
Use LLM to parse free text synthesis recipes to structured data! With only 300 training pairs, the fine tuned model can already pick up chemical identities/amounts and generate valid JSON in ORD schema. Check out our video demo!
1
3
19
Proud to be part of this massive #llmhack manuscript invited by none other than my Ph.D. advisor @A_Aspuru_Guzik, and to be part of the pioneering movement to incorporate LLMs into how we solve chemical problems!.
11. @QaiAlex, @jacosand, @Jiale_Shi, Stefan Bringuier, @BrendenPelkie, Marcus Schwarting use LLMs to convert procedure texts in the ORD to structured JSON.
0
0
5
Enjoy the #llmhack!.
Thanks to @kmjablonka for giving a lunchtime talk about the opportunities for LLMs in Chemistry.
0
0
6
@QaiAlex The model was tuned on data from the Open Reaction Database led by @cwcoley, Abby Doyle, and many others. Thanks to @BenBlaiszik, @kmjablonka, and @andrewwhite01 for organizing #llmhack! Hope to continue working on improving the LLM!.
2
0
4
I really loved reading the summary of this project! Incredible what the gptchem approach can do. Also, great work making it run on consumer hardware!. Look out for a summary of all the #llmhack projects coming out soon!.
🧪 Worked on a new project w/ Ankur Gupta and @merz_garrett using LLMs to speed up accurate prediction of chemical properties. To start, we use GPT-3 with GPTChem & GPT2-LoRA to predict the atomization energies of molecules at the G4(MP2) theory level from the QM9 dataset.
1
3
17
Wow, natural language interface and helper prototype built for 3dmol.js. Reach out to @jakublala if you are interested! 🤩. #llmhack.
As a part of the #llmhack, me and Sean Warren developed MolVerse, a language interface to 3dmol.js: a web app for all chemists and biologists that want to visualise their (bio)molecules without needing to code. Here we load the structure of benzene:
0
1
11
Still blows my mind what you can hack together in a few hours! . It was a pleasure seeing all the participants having so much (productive!) fun with LLMs in materials and chemistry! #llmhack
We wrapped up the first LLM hackathon for applications in materials and chemistry last week. The results to me were astounding. We are at the point now where some tasks that took years can now be completed in days. Here is a list of the fantastic submissions!
0
2
9