According to emollick on X, creator details surfaced via RyanMorey showing a small LLM called Mr. Chatterbox trained end-to-end with Andrej Karpathy’s Nanochat on Victorian-era books (1837–1899), using a subset of the BL Books dataset and two rounds of supervised fine-tuning to handle style fidelity and safety edge cases (source: Ethan Mollick on X; Ryan Morey on X; Nanochat GitHub discussions). According to RyanMorey, the pipeline used Nanochat for initial training and SFT, with round one covering 2 epochs over 40,000+ corpus and synthetic pairs, and a second focused round for modern greetings, goodbyes, and prompt-injection defense, indicating practical methods for domain-style alignment and guardrail tuning in small models (source: Ryan Morey on X; Nanochat GitHub discussions). As reported by Ethan Mollick, this demonstrates a low-cost approach for enterprises to build brand-voice assistants and historical-domain chatbots by combining curated domain corpora with targeted SFT, suggesting opportunities for boutique LLMs in publishing, museums, education, and heritage tourism (source: Ethan Mollick on X).

According to DeepLearning.AI, most large language models struggle with tasks like counting specific letters in words due to tokenizer limitations and inadequate evaluation methods. In the course 'Fine-tuning and Reinforcement Learning for LLMs: Intro to Post-Training' taught by Sharon Zhou, practical techniques are demonstrated for designing evaluation metrics that identify such issues. The course also explores how post-training approaches, including supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF), can guide models toward more accurate and desirable behaviors, addressing real-world application challenges for enterprise AI deployments. As reported by DeepLearning.AI, these insights empower practitioners to improve LLM performance through targeted post-training strategies.