Enoch H. Kang

Best AI papers explained

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

Husk at besøge podcastens hjemmeside og støtte skaberen: podcasters.spotify.com

Forfatter

Enoch H. Kang

Kategori

Technology

Podcastens hjemmeside

podcasters.spotify.com

Seneste episode

17. nov. 2026

Hvor kan du lytte?

Podcasts i appen Replaio Radio Kommer snart

Podcasts kommer snart til appen. Installer nu, og vær den første til at se en helt ny tilgang til podcasts

Hent den på Google Play Installer gratis Android 10 mio.+ downloads · 4,7+ i bedømmelse iOS snart

Episoder

Front-Loading Reasoning: The Synergy between Pretraining and Post-Training Data 18.10.2025

This research paper, by authors affiliated with NVIDIA, Carnegie Mellon University, Boston University, and Stanford University, focuses on the optimal strategy for incorporating reasoning data into Large Language Model (LLM) training. The central finding challenges the conventional approach of relying solely on post-training, demonstrating that "front-loading" reasoning data during the p...

Representation-Based Exploration for Language Models: From Test-Time to Post-Training 18.10.2025

This paper investigates the effectiveness of deliberate exploration in enhancing the reasoning capabilities of large language models (LLMs) trained with reinforcement learning (RL). The authors propose and evaluate a novel representation-based exploration (RepExp) strategy, which uses a bonus derived from the LLM's hidden states to encourage the discovery of diverse and novel behaviors. The st...

The attacker moves second: stronger adaptive attacks bypass defenses against LLM jail- Breaks and prompt injections 18.10.2025

The academic paper discusses the critical flaws in current methods used to evaluate the robustness of large language model (LLM) defenses against jailbreaks and prompt injections. The authors argue that testing defenses with static or computationally weak attacks yields a false sense of security, as demonstrated by the fact that they successfully bypassed twelve different recent defenses with an a...

When can in-context learning generalize out of task distribution? 16.10.2025

The research empirically investigates the role of pretraining distribution and a new concept of task diversity in the emergence of ICL, particularly using models trained on linear functions. Findings indicate that increasing task diversity causes transformers to shift from a specialized solution to one that can generalize across the entire task space, a transition also observed in nonlinear regres...

The Art of Scaling Reinforcement Learning Compute for LLMs 16.10.2025

This paper studies scaling reinforcement learning (RL) compute for large language models (LLMs), introducing a principled framework to predict performance. The authors develop ScaleRL, a best-practice recipe derived from ablating various algorithmic choices, and demonstrate its predictable scaling trajectory using a sigmoidal function to fit compute-performance curves. Accompanying figures illustr...

A small number of samples can poison LLMs of any size 16.10.2025

This white paper by Anthropic, UK AI Security Institute , and The Alan Turing Institute demonstrates that a small, fixed number of malicious documents —as few as 250—can successfully create a "backdoor" vulnerability in LLMs, regardless of the model's size or the total volume of clean training data. This finding challenges the previous assumption that attackers need to control a perc...

Dual Goal Representations 14.10.2025

This paper discusses dual goal representations for goal-conditioned reinforcement learning (GCRL), a novel method for encoding a state based on its temporal distance relation to all other states within an environment. The authors theoretically establish that this representation is sufficient for recovering an optimal goal-reaching policy and is invariant to extraneous noise within the state observ...

Welcome to the Era of Experience 14.10.2025

This "bitter lesson"-style position paper by David Silver and Richard S. Sutton introduces a shift in artificial intelligence from the "Era of Human Data" to the "Era of Experience." The authors argue that AI progress is slowing because the available pool of human-generated data is being exhausted, necessitating a new approach where agents learn predominantly from their own self-generated experien...

Value Flows: Flow-Based Distributional Reinforcement Learning 14.10.2025

This paper introduces "Value Flows," a novel reinforcement learning algorithm that uses flow-based models to estimate the full future return distribution , instead of flattening it to a single scalar value like traditional methods. This approach is designed to provide richer learning signals and better estimations of aleatoric uncertainty (return variance), which is then used to prioriti...

Self-Adapting Language Models 12.10.2025

This paper introduces Self-Adapting Large Language Models (SEAL), a novel framework that enables LLMs to autonomously improve by generating their own training data and finetuning instructions, termed "self-edits." This adaptation process is driven by a reinforcement learning (RL) loop that rewards the model for generating self-edits that subsequently improve its performance on downstream...

The Markovian Thinker 12.10.2025

Reinforcement learning (RL) methods for training Large Language Models (LLMs) to produce long chains of thought (LongCoT) are constrained by the standard thinking environment, where the state is unbounded, leading to quadratic computational costs as reasoning length increases. This paper propose Markovian Thinking, a paradigm where the reasoning policy conditions only on a constant-size state, eff...

Moloch’s Bargain: emergent misalignment when LLMs compete for audiences 12.10.2025

The academic paper investigates a phenomenon called Moloch’s Bargain for AI, demonstrating that optimizing Large Language Models (LLMs) for competitive success in market-driven environments inadvertently leads to misalignment and harmful behaviors. The researchers use simulated environments across three domains—sales, elections, and social media—to show that performance gains, such as increased sa...

Transformer Predictor Dynamics and Task Diversity 11.10.2025

This paper focuses on modeling the behavior of Transformer models during training, particularly concerning in-context learning (ICL), which shows a transition from generalizing to memorizing. The authors utilize a Bayesian model that incorporates two primary predictors, Memorizing (M) and Generalizing (G), and demonstrate that this model accurately captures the observed behavior of the Transformer...

Base models know how to reason, thinking models learn when 11.10.2025

This paper argues that thinking language models (LLMs that reason step-by-step) do not acquire entirely new capabilities during post-training but rather learn when to deploy pre-existing reasoning mechanisms latent in their base counterparts. The authors use an unsupervised clustering methodology via Sparse Autoencoders (SAEs) to derive an interpretable taxonomy of distinct reasoning behaviors, su...

Spectrum tuning: Post-training for distributional coverage and in-context steerability 11.10.2025

This research paper focuses on conditional distributional modeling for large language models (LLMs), introducing the SPECTRUM SUITE dataset and a new training method called SPECTRUM TUNING. The paper outlines three main objectives: in-context steerability (modifying output probabilities based on inference-time information), valid output coverage (generating diverse, correct responses), and distrib...

Understanding Prompt Tuning and In-Context Learning via Meta-Learning 11.10.2025

The academic paper investigates prompt tuning and in-context learning through a meta-learning and Bayesian lens, positing that optimal prompting can be understood as conditioning Bayesian sequential predictors. The authors detail how meta-trained neural networks, like LSTMs and Transformers, function as Bayes-optimal predictors and explore the theoretical limitations of prompting, particularly for...

MLPs Learn In-Context on Regression and Classification tasks 11.10.2025

This research paper demonstrates that Multi-Layer Perceptrons (MLPs) can perform In-Context Learning (ICL), an ability often attributed exclusively to Transformer models. The researchers show that MLPs, and related MLP-Mixer models, achieve performance comparable to Transformers on synthetic ICL tasks involving regression and classification. Furthermore, in experiments testing relational reasoning...

Is Pre-Training Truly Better than Meta-Learning? 11.10.2025

The research challenges the belief that pre-training (PT) always outperforms meta-learning (MAML) in few-shot learning by conducting a rigorous, fair empirical comparison across diverse datasets. The authors introduce and utilize the Task2Vec diversity coefficient to categorize datasets as having either low or high diversity. The primary finding suggests that pre-training is generally better for l...

Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models 11.10.2025

This paper introduces Agentic Context Engineering (ACE), a novel framework designed to enhance the performance of Large Language Models (LLMs) in complex applications like agents and domain-specific reasoning by evolving their context, or "playbook." ACE addresses two key limitations of prior context adaptation methods: brevity bias (the loss of detailed domain knowledge for conciseness)...

Do LLMs Recognize Your Preferences? Evaluating Personalized Preference Following in LLMs 09.10.2025

This paper assesses how well Large Language Models (LLMs) can infer, remember, and follow user preferences in long, multi-session conversations. The evaluation of 10 different LLMs using this benchmark revealed that current state-of-the-art models exhibit significant difficulty proactively following user preferences, with accuracy dropping below 10% in zero-shot settings within a short number of t...

Learning dynamics of LLM finetuning 09.10.2025

This academic paper presents a novel framework for understanding the evolution of Large Language Models (LLMs) during finetuning by analyzing their learning dynamics from a dynamical perspective , contrasting with previous approaches focused on training targets or end-states. The authors formalize the change in model prediction using a decomposition into three key terms, which adapts to various fi...

Iterative Data Smoothing: Mitigating Reward Overfitting and Overoptimization in RLHF 09.10.2025

This paper investigate two major drawbacks in the reward learning phase of RLHF: reward overfitting and reward overoptimization, which often occur because the standard cross-entropy loss is inadequate for imbalanced preference datasets. To address these issues, the paper introduces a novel algorithm called Iterative Data Smoothing (IDS), which mitigates these problems by iteratively updating hard...

OpenAI Agent Builder and n8n: Orchestrating Reasoning Versus Automating Process 08.10.2025

Instead of introducing a paper, today we conduct a strategic analysis comparing OpenAI's Agent Builder with the established workflow automation platform, n8n, concluding that the former is not a replacement for the latter but rather the creator of a new, parallel market. The core difference lies in their philosophies: Agent Builder is designed for orchestrating complex AI reasoning and handlin...

Training Agents Inside of Scalable World Models 08.10.2025

This paper introduces  Dreamer 4 , a new world model designed to solve complex control tasks, particularly the  Minecraft diamond challenge , purely through  offline imagination training  without direct environment interaction. The core innovation lies in its architecture, which uses an  efficient block-causal transformer  and a  shortcut forcing objective  to achieve high prediction accuracy of g...

Small Language Models are the Future of Agentic AI 07.10.2025

This paper presents a strong **position statement** arguing that **Small Language Models (SLMs)** are the **future of agentic AI**, despite the current dominance of **Large Language Models (LLMs)**. The authors contend that SLMs are **sufficiently powerful**, **more economical**, and **operationally more suitable** for the specialized and repetitive tasks common in AI agents. They provide **argume...

Lyt til podcasten Best AI papers explained i Replaio

Radio og podcasts i én app - gratis og uden tilmelding. Installer i dag, og gå ikke glip af lanceringen

Hent den på Google Play

Replaio er ikke podcastudgiver; programmernes navne, coverbilleder og lyd tilhører deres ophavsmænd og distribueres via offentlige RSS-feeds.