Loading…
Hosted by Unknown Host · EN · 45 episodes
Established thought leaders with verified media credentials.
A conversational roadmap for the full language-model series, introducing the expert mental models and major disagreements that connect architecture, scaling, optimization, fine-tuning, RLHF, open models, and sparse experts.
Unknown Host hosts Mastering Language Models: From Architecture to Optimization, a general show with 45 episodes published.
In the finale of Topic 7 and the series, Maya and Leo unpack a 2025 systematic review of optimization methods for machine learning — the optimizer that decides whether a ninety-day run converges. They map the gradient-ba
Maya and Leo trace AutoML from grid and random search through Bayesian optimization, neural architecture search, and automated pipeline construction (TPOT, AutoGluon, Auto-Keras), then ask where automated search beats ex
The Pile reframes training data as a design choice. EleutherAI assembled an 825-gibibyte English corpus from twenty-two diverse sub-datasets — books, code, biomedical and physics papers, legal opinions, dialogue, and a f
Maya and Leo follow the sparsely-gated Mixture-of-Experts idea out of language and into scientific machine learning. The paper, MoE-POT, is an operator transformer pre-trained across many families of partial differential
Maya and Leo unpack the 2017 paper that started sparse scaling. They walk from the wall it was hitting — capacity and compute welded together in dense models — to the central move: a Mixture-of-Experts layer where a nois
Maya and Leo open Topic 7 by going underneath massive models to the machinery that makes them possible. Using one small lab training a large sparse model on a fixed ninety-day compute budget, they explain sparsity and Mi
Maya and Leo unpack DeepSeek-V4-Flash as an efficiency stack for million-token context: hybrid compressed attention, sparse expert activation, low-precision serving, and specialist distillation. They stage the field'
This episode uses John Schulman's Berkeley EECS colloquium on RLHF progress and challenges to explain why human preference training made assistants more useful, and why reward proxies, reviewer disagreement, halluci
A deep dive into VRPO, a robust RLHF fine-tuning method that uses auxiliary preference models and reference-policy samples to reduce noisy preference-objective variance while preserving the need for domain-specific safet
A deep dive into the critical RLHF Deciphered paper, explaining why reward models are useful but fragile proxies, how sparse human feedback can misgeneralize, and what production teams should audit before trusting prefer
Maya and Leo explain Direct Preference Optimization, showing how DPO turns preference pairs into a direct language-model update that behaves like an implicit reward model while avoiding a separate reward-model-plus-RL pi
Maya and Leo explain Constitutional AI as a shift from many direct human harmlessness labels to written principles that guide model critiques, revisions, and AI-generated preference comparisons. The episode uses a health
Maya and Leo unpack Anthropic's helpful-and-harmless RLHF paper: how comparison data trains a preference model, how RL steers assistant behavior, why helpfulness and harmlessness can fight each other, and why reward
This episode explains the InstructGPT paper's core training recipe: supervised demonstrations, human preference rankings, reward modeling, and careful RLHF optimization. Maya and Leo use a healthcare scheduling assi
Maya and Leo unpack the OpenAI paper Learning to Summarize from Human Feedback, showing how pairwise human preferences, reward modeling, and PPO moved summarization beyond reference matching while introducing new risks a
Maya and Leo explain Proximal Policy Optimization as the controlled update mechanism inside many RLHF systems. Using a healthcare scheduling assistant as the running example, they unpack old-policy anchoring, clipped pro
A topic overview of RLHF: how human comparisons become preference data, how reward models and cautious optimization steer assistant behavior, why PPO and DPO represent different engineering philosophies, and where feedba
Updating models without letting yesterday’s skills disappear. Sources: • Continual Learning of Large Language Models: A Comprehensive Survey: https://arxiv.org/abs/2404.16789
How far can precision shrink before adaptation breaks?. Sources: • LowRA: Accurate and Efficient LoRA Fine-Tuning of LLMs under 2 Bits: https://arxiv.org/abs/2502.08141
Training adapters through a frozen four-bit model. Sources: • QLoRA: Efficient Finetuning of Quantized LLMs: https://arxiv.org/abs/2305.14314
Sponsor detection runs nightly. Check back soon.
No public pitch examples yet for this show.
Generate your own personalised pitchBased on semantic analysis of episode topics and host coverage, this show is a strong guest fit for executives in:
Industry fit is computed by PitchCentric using vector embeddings of the show's episode catalog.
Shows with the most semantically similar episode content. Pitch one, pitch all; producers cluster.
Mastering Language Models: From Architecture to Optimization has a verified contact on file. Create a free PitchCentric account to access it and generate a personalised pitch in seconds. Research at least 3 recent episodes first and lead with a specific angle that serves their general audience.
Mastering Language Models: From Architecture to Optimization is hosted by Unknown Host. The show is categorised under General and has published 45 episodes.
Mastering Language Models: From Architecture to Optimization has published 45 episodes.
Mastering Language Models: From Architecture to Optimization is accessible for guests with genuine general expertise. A personalised, episode-aware pitch will still outperform a generic one every time.
Mastering Language Models: From Architecture to Optimization hasn't explicitly signalled guest openness in recent episodes. That doesn't rule out pitching. your hook just needs to be especially compelling and relevant to their recent content.
Episodes of Mastering Language Models: From Architecture to Optimization average 13 minutes. a focused format where a clear narrative arc and tight preparation matter most.
Our data rates Mastering Language Models: From Architecture to Optimization's guest bar at 80/100 (Premium tier). Established thought leaders with verified media credentials. Sign in to PitchCentric to see how your own Pod Score compares against this show.
Methodology. Booking Probability™ blends Listen Score, 30-day Virality, open-to-guests detection, and Apple ratings. Data refreshed every 60 minutes. Listen Score and Booking Probability are calculated by PitchCentric. Last enriched 2 days ago.