1. Podcasts
  2. Technology
  3. Best AI papers explained
Best AI papers explained cover art

530 episodes

Best AI papers explained, page 17

by Enoch H. Kang · English

Cut through the noise. We curate and break down the most important AI papers so you don’t have to.

All episodes, page 17

Self-Adapting Language Models

Self-adapting language models (SEAL) that improve by generating “self-edits” (natural-language instructions) to create their own training data and update their weights, using meta-learning.

12 Oct 2025 · 17m · 8 chapters

The Markovian Thinker

The episode explains why “long chain of thought” (long-CoT) training is so expensive for LLMs (quadratic compute from transformer attention over growing context, plus linear KV-cache memory growth),…

12 Oct 2025 · 14m · 6 chapters

Moloch’s Bargain: emergent misalignment when LLMs compete for audiences

The episode explains “Moloch’s bargain” in AI: when LLMs are optimized to win against competitors for short-term audience approval (sales, votes, engagement), long-term goals like truth and safety…

12 Oct 2025 · 17m · 12 chapters

Transformer Predictor Dynamics and Task Diversity

Explains in-context learning (ICL) variability using a top-down “rational analysis” framework: a hierarchical Bayesian model where transformer behavior is a weighted mix of a generalizing predictor…

11 Oct 2025 · 16m · 9 chapters

Base models know how to reason, thinking models learn when

Why “thinking models” (e.g., DeepSeek R1, Claude 3.7 Sonnet) outperform base LLMs on hard math—arguing that base models already contain core reasoning skills, and post-training mainly teaches…

11 Oct 2025 · 12m · 6 chapters

Spectrum tuning: Post-training for distributional coverage and in-context steerability

How post-training instruction tuning can damage a model’s conditional distributional modeling (CDM), reducing in-context steerability, output diversity/coverage, and distributional alignment; and how…

11 Oct 2025 · 16m · 10 chapters

Understanding Prompt Tuning and In-Context Learning via Meta-Learning

Explains why LLM prompting works using a Bayesian meta-learning view of in-context learning, and compares hard prompt tuning vs prefix tuning (soft prompts), including when prompting fails and how…

11 Oct 2025 · 14m · 6 chapters

MLPs Learn In-Context on Regression and Classification tasks

Compares in-context learning (ICL) across MLPs, MLP mixers, and transformers on synthetic regression/classification and relational reasoning tasks, focusing on compute efficiency, context length…

11 Oct 2025 · 16m · 10 chapters

Is Pre-Training Truly Better than Meta-Learning?

Few-shot learning comparison between pre-training (PT) + head fine-tuning and model-agnostic meta-learning (MAML), arguing PT isn’t universally better; results depend on task/data diversity measured…

11 Oct 2025 · 21m · 11 chapters

Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models

Agentic Context Engineering (ACE) for self-improving LLM agents—how to adapt task context over time without catastrophic forgetting or “context collapse,” using structured, incremental updates and…

11 Oct 2025 · 18m · 9 chapters

Do LLMs Recognize Your Preferences? Evaluating Personalized Preference Following in LLMs

The episode discusses research on whether large language models can recognize and consistently follow user preferences over long conversations, and why they often “forget” (the “zero-shot crisis”).

9 Oct 2025 · 16m · 9 chapters

Learning dynamics of LLM finetuning

Learning dynamics in LLM fine-tuning—how SFT and DPO change predictions across unrelated examples, causing hallucinations and repetition via an ENTK-driven “squeezing effect.”

9 Oct 2025 · 12m · 7 chapters

Iterative Data Smoothing: Mitigating Reward Overfitting and Overoptimization in RLHF

RLHF alignment failures caused by reward overfitting (reward model generalization worsens after ~1 epoch) and reward overoptimization (policy exploits a flawed reward proxy, reducing ground-truth…

9 Oct 2025 · 17m · 7 chapters

OpenAI Agent Builder and n8n: Orchestrating Reasoning Versus Automating Process

Compares OpenAI Agent Builder (AgentKit) vs n8n for enterprise workflows, framing a “reasoning vs orchestration” split and arguing for a hybrid stack.

8 Oct 2025 · 15m · 9 chapters

Training Agents Inside of Scalable World Models

Google DeepMind’s Dreamer 4 trains agents inside scalable world models using offline “imagination” in Minecraft, solving the diamond-finding challenge as a 20,000+ action, pixel-observation…

8 Oct 2025 · 14m · 10 chapters

Small Language Models are the Future of Agentic AI

The episode argues that agentic AI should shift from relying on centralized large language models (LLMs) to an “SLM-first” architecture using small language models (under ~10B parameters) for most…

7 Oct 2025 · 19m · 9 chapters

Activation Steering in Generative Settings via Contrastive Causal Mediation Analysis

Contrastive Causal Mediation Analysis (CCM) for precisely locating and steering internal LLM components to control freeform generative behaviors, using contrastive prompt/response pairs and efficient…

6 Oct 2025 · 18m · 9 chapters

Eliciting Secret Knowledge from Language Models

The episode covers “secret elicitation” in language models—how models can be fine-tuned to conceal knowledge they still use, and how researchers audit/expose that concealment using black-box and…

6 Oct 2025 · 15m · 7 chapters

Temporal difference flow

Long-horizon reinforcement learning for predictive AI, focusing on temporal difference flows (TD flow) to train geometric horizon models (GHMs) without unstable, noisy learning.

6 Oct 2025 · 15m · 8 chapters

Personalized reasoning: just-in-time personalization and why LLMs fail at it

The episode argues that LLMs often get “correctness” but fail “just-in-time personalization,” because training typically separates factual accuracy from preference alignment.

5 Oct 2025 · 14m · 5 chapters

Prompt Curriculum Learning for Efficient LLM Post-Training

Prompt Curriculum Learning (PCL) for efficient LLM post-training, aiming to improve reasoning (e.g., math) without the heavy cost of RL with verifiable rewards (RLVR).

5 Oct 2025 · 13m · 6 chapters

Personalizing Reinforcement Learning from Human Feedback with Variational Preference Learning

Pluralistic alignment for AI systems—personalizing reinforcement learning from human feedback (RLHF) so models don’t average conflicting user preferences.

4 Oct 2025 · 18m · 13 chapters

Enhancing Personalized Multi-Turn Dialogue with Curiosity Reward

Curiosity-driven user modeling reward (Cure.io) for personalized multi-turn dialogue, aiming to avoid “one-size-fits-all” recommendations by giving LLMs dense, intrinsic feedback during chats.

4 Oct 2025 · 14m · 7 chapters

Learning to summarize user information for personalized reinforcement learning from human feedback

PLUS (pluralistic alignment via summarization) improves RLHF personalization for diverse users by replacing a single-average Bradley-Terry-Luce reward model with user-specific, text-based preference…

4 Oct 2025 · 16m · 8 chapters