Leonard Salewski

Aleph Alpha

@L_Salewski

Post-training LLMs at @Aleph__Alpha | Ex intern at @Nvidia and @Bosch_AI

Santa Clara Joined May 2023

200 Following 212 Followers 89 Posts

Latest posts 5

Pinned post

Leonard Salewski @L_Salewski Jul 15, 2023

📢 #scholarGPT now can perform complex filtering operations like selecting a specific year 📅, making research even more efficient! 💪 Build with self-query from @langchain
Sign up on the waitlist to be the first to try it out mailchi.mp/d8b3dbb881ae/s…! 🚀✨ #ChatGPT #Research🧑‍💻

0:13

Leonard Salewski @L_Salewski Nov 3

One fun aspect of our latest blog post: You can change its reasoning effort.

Drag the slider and the post rewrites itself word by word, from a 12-minute read at high down to 2 minutes at none.

0:15

Aleph Alpha @Aleph__Alpha Nov 3

Kolibri can reason at four efforts: none, low, medium, and high. Our SFT training traces didn't have effort labels, so we had to label them post-hoc.

A new blog post by @L_Salewski explains how we did this, with interactive figures 🧵

View quoted post on X

Leonard Salewski @L_Salewski Oct 26

The Kolibri tech report is now on arXiv: arxiv.org/abs/2610.31945. If you haven't read it yet, give it a read - it covers all steps on how to build a modern LLM.

Screenshot of the cover page of the Kolibri Tech Report uploaded to arXiv. Contains the abstract and Figure 1.

Leonard Salewski @L_Salewski Oct 3

Here are two neat speed tricks we found while SFT'ing Kolibri 🧵

1: We use standard Next-k-fit online bin packing with document masking maximises useful tokens per sequence. Whats easy to miss: less padding means a larger effective batchsize, so the learning rate can go up.

0:08

Aleph Alpha @Aleph__Alpha Oct 3

In post-training, we taught Kolibri to think and answer in both English and German consistently across SFT and RL alike. We generated large amounts of synthetic data to control how Kolibri answers and trained with RL on more than 1.2 million tasks, including ones built around our customer needs, and ones to teach it to say “I don’t know” when the context is insufficient.

View quoted post on X

Leonard Salewski @L_Salewski Oct 3

Here is one interesting detail from our post-training:

Kolibri scores above 90% on AIME 2025 and AIME 2026 just after SFT. We also evaluated on an internal AIME 2026-like eval with new IID samples and got even better results. For math, RL was just the finishing touch than.

Aleph Alpha @Aleph__Alpha Oct 3

In post-training, we taught Kolibri to think and answer in both English and German consistently across SFT and RL alike. We generated large amounts of synthetic data to control how Kolibri answers and trained with RL on more than 1.2 million tasks, including ones built around our customer needs, and ones to teach it to say “I don’t know” when the context is insufficient.

View quoted post on X