Research
Announcing Talkie: a new, open-weight historical LLM! We trained and finetuned a 13B model on a newly-curated dataset of only pre-1930 data.…
Talkie is a new 13-billion parameter open-weight language model trained exclusively on pre-1930 historical data through a newly-curated dataset. The model was trained and fine-tuned to specialize in u
Talkie is a new 13-billion parameter open-weight language model trained exclusively on pre-1930 historical data through a newly-curated dataset. The model was trained and fine-tuned to specialize in understanding and generating text from the historical period before 1930. This development aims to provide researchers and developers with an AI tool optimized for historical text analysis and generation.
Related
- Test of Time LLM from some legends...pretty fun!
- REALM: Reliable Expertise-Aware Language Model Fine-Tuning from Noisy Annotations
- - Hermes Trismegistus
- New work with @AlecRad and @DavidDuvenaud: Have you ever dreamed of talking to someone from the past? Introducing talkie, a 13B model traine…
Source: Soumith Chintala (X) | 2026-04-27