Skip to content
English

Switch theme

Discover▾

Not sure where to start? This week’s hottest software, models and stacks in three minutes.

AI Apps▾

AI software people use: from assistants like ChatGPT to tools for coding, video and workflows, picked by purpose.

For Agents▾

Software for AI: tools that let your agent (Claude Code, Codex, Cursor…) browse, read docs, run code and remember.

Models▾

Choose models, estimate cost, self-host: prices, rankings, benchmarks, free tiers and compute.

Solutions▾

A goal or an error message? Ready-made stacks, steps and fixes by scenario.

Learn▾

A systematic course from zero to the frontier, with an algorithm library, must-read papers and tutorials.

Frontier▾

Daily AI news, weekly highlights, the latest papers and how AI is changing every discipline.

// Course · stage 9 of 9

Safety, Evaluation & Frontier

Evaluation and benchmarks, interpretability, security, governance, AI for Science and keeping up with the frontier.

Start the first unit → All stages

6 unitsabout 44 hours29 algorithms

01Units in this stage

In order. Each unit has goals, algorithms, must-read papers, a project and a self-test.

01 · U90 Evaluation & Benchmarks Major benchmarks, arenas, data contamination and the limits of capability evaluation. 6 h2 algorithms2 paperslesson ready 02 · U91 Interpretability: from Attribution to Circuits Attribution, probes, the logit lens, superposition, SAEs, circuit tracing, knowledge editing and activation steering. 10 h9 algorithms4 paperslesson ready 03 · U92 Security: Jailbreaks, Injection & Privacy Adversarial examples, GCG, prompt injection, red-teaming, watermarking, differential privacy and backdoors. 6 h8 algorithms4 paperslesson ready 04 · U93 Frontier Risk, Scalable Oversight & Governance Weak-to-strong generalisation, debate, Constitutional AI, risk frameworks and regulation (EU AI Act, NIST AI RMF). 6 h4 algorithms3 papers 05 · U94 AI for Science & Domain Applications Graph neural networks, AlphaFold, AlphaGeometry and automated research — linked with AIMap’s “AI × Disciplines” section. 10 h4 algorithms3 papers 06 · U95 Tracking the Frontier: Reading & Reproducing Papers The three-pass reading method, tracking arXiv and conferences, reproduction and writing. 6 h3 algorithms0 papers

02Algorithms you will meet

Algorithms Atlas →

LLM-as-a-JudgeEmergent AbilitiesFeature Attribution (Integrated Gradients / SHAP)Grad-CAMLogit Lens & ProbingCircuits & Induction HeadsSuperpositionSparse Autoencoders (SAE)Circuit Tracing & Attribution GraphsLocating & Editing Facts (ROME)Activation SteeringAdversarial Examples (FGSM)GCG Universal Jailbreak AttackIndirect Prompt InjectionAutomated Red TeamingLLM WatermarkingDifferentially Private SGD (DP-SGD)Sleeper Agents (Backdoors that Persist)Membership Inference AttacksWeak-to-Strong GeneralizationAI Safety via DebateConstitutional AI / RLAIFGraph Neural Networks (GCN / GAT)AlphaFold 2AlphaGeometry (Neuro-Symbolic Reasoning)RL with Verifiable Rewards (RLVR)Transformer & Self-AttentionGroup Relative Policy Optimization (GRPO)Flow Matching

Part of these routes: Everyday Users · Application Developers · Algorithm Researchers · Product Managers & Leaders · Domain Scientists

Data updated: AI algorithms & curriculum(UTC+8)