Index additions: 2026-W42
Resources added to the index in this ISO week. This is a website archive, not a sent newsletter.
141 resources · page 4 of 5
Model Training
Reinforcement Learning from Human Feedback (book)
A conceptual path through reward modeling and post-training methods.
Controlling Reasoning Effort in LLMs (Ahead of AI)
A free Ahead of AI article on training controllable reasoning effort.
Unsloth
by Unsloth
Study local training and adaptation workflows for supported models.
nanochat
A compact training harness spanning the model-development lifecycle.
DeepSeek-R1: Incentivizing Reasoning via RL (arXiv 2501.12948)
by DeepSeek
Study the reported reinforcement-learning recipe behind DeepSeek-R1.
The Smol Training Playbook
A training account that includes failed experiments and engineering decisions.
The Ultra-Scale Playbook
Understand how parallelism changes memory use and training throughput.
The FineWeb Datasets
by Guilherme Penedo and coauthors, Hugging Face
Study how filtering and deduplication affect pretraining data.
verl (HybridFlow)
by verl project (originated at ByteDance Seed)
Explore how an RL training framework coordinates generation and optimization.
Jailbreaking and red teaming
Indirect prompt injection in the real world: how people manipulate neural networks
by Vladislav Tushkanov (Kaspersky)
Read a Russia-origin security team’s field study of indirect prompt injection and the implications for defensive risk assessment.
NIST AI Risk Management Framework
by NIST
A government reference for organizing AI risk-management work.
OWASP Top 10 for LLM Applications 2025
by OWASP Gen AI Security Project
Use a shared vocabulary for discussing risks in LLM applications.
Simon Willison on prompt injection
Follow an ongoing public record of prompt-injection research.
The lethal trifecta for AI agents
A defensive framework for recognizing risky combinations of agent capabilities and untrusted inputs.
Agentic misalignment
by Anthropic
A controlled study of risky behavior under conflicting agent incentives.
JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models
by Patrick Chao, Edoardo Debenedetti, Alexander Robey, Maksym Andriushchenko and colleagues
Learn why jailbreak studies need consistent threat models, scoring and reproducible comparisons.
Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection
Foundational public research on indirect prompt injection. Linked only; no payloads reproduced.
Embrace The Red
Public case reports about security failures in deployed AI products.
LLAMATOR: authorized conversational AI security testing
by Roman Neronov, Timur Nizamov and Nikita Ivanov / LLAMATOR
A Russian-origin framework for assessing your own chatbots, with reviewable test histories and reports.
HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
by Mantas Mazeika, Long Phan and colleagues
Study a standardized approach to red-team evaluation and robust refusal without copying attack recipes.
MITRE ATLAS
by MITRE
Connect AI threats with documented techniques, mitigations and cases.
garak - LLM vulnerability scanner
by NVIDIA
A public scanner project for understanding classes of model failures.
promptfoo
by promptfoo
Inspect a project combining application evaluation and security testing.
Qwen3Guard Technical Report
by Haiquan Zhao and colleagues / Qwen Team
Understand Qwen’s multilingual safety classifiers and the distinction between whole-response and streaming moderation.
Design Patterns for Securing LLM Agents against Prompt Injections
Summary of a paper's defensive agent architectures.
Frontier
Responsible Scaling Policy
by Anthropic
Read a lab’s published risk-governance commitments and revision history.
Artificial Analysis
Compare model and provider characteristics in one public analysis hub.
Epoch AI - Data on the trajectory of AI
by Epoch AI
Explore longitudinal datasets about models, compute and infrastructure.
Hugging Face Daily Papers
by Hugging Face
A community-curated route into newly discussed AI papers.
Critiquing Anthropic’s Variant of Pacing the Frontier (ChinAI)
by Jeffrey Ding
A free ChinAI essay examining arguments about international AI competition.
State of AI Report 2026
by Nathan Benaich / Air Street Capital
A yearly synthesis of AI research, industry and policy developments.
The current balance of power in open models (Interconnects)
A free Interconnects essay distinguishing open-weight and closed-model ecosystems.