Safety Benchmark WMDP
The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning Nathaniel Li, Alexander Pan, Anjali Gopal, Summer Yue, Daniel Berrios, Alice Gatt...
Posts by Tag
- Adversarial 9
- Safety 9
- Efficiency 8
- LLMEvaluate 8
- BasicLLM 7
- Agent 6
- Mitigate 5
- Applications 5
- RL 3
- AGI 3
- language model 3
- RAG 3
- Reasoning 3
- Train 3
- Customization 3
- Alignment 2
- Prompting 2
- Jailbreaking 2
- Serving 2
- Scaling 2
- Protein 1
- Diffusion 1
- Image synthesis 1
- Human Alignment 1
- Bias 1
- Hallucination 1
- DomainAdapt 1
- ModelEdit 1
- Interpretibility 1
- LongContext 1
- Planning 1
- Multiagent 1
- Multimodal 1
Adversarial
More FM risk
In this session, our readings cover:
LLM multimodal harm responses
In this session, our readings cover:
FM toxicity / harmful outputs
In this session, our readings cover:
FM fairness / bias issues
In this session, our readings cover:
FM privacy leakage issues
In this session, our readings cover:
FM copyright infrigement
In this session, our readings cover:
Survey AI Risk framework
In this session, our readings cover:
GenAI Guardrails
In this session, our readings cover:
Safety
Model Interpretibility for FM
In this session, our readings cover:
Agent Safety
In this session, our readings cover:
Extra readings - Agent Guardrailing
In this session, our readings cover:
Platform - VLM Jailbreaking / Probing
In this session, our readings cover:
Platform - Model Jailbreaking / Safeguarding
In this session, our readings cover:
Safety Benchmark WMDP
The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning Nathaniel Li, Alexander Pan, Anjali Gopal, Summer Yue, Daniel Berrios, Alice Gatt...
More FM risk
In this session, our readings cover:
LLM multimodal harm responses
In this session, our readings cover:
FM toxicity / harmful outputs
In this session, our readings cover:
Efficiency
Privacy
Beyond Memorization: Violating Privacy Via Inference with Large Language Models
Gradient for LLM
TextGrad: Automatic “Differentiation” via Text
KV Cache and Tooling
KV Caching in LLM:
Advanced Transformer Architectures
In this session, our readings cover:
Recent LLM basics
In this session, our readings cover:
LLM Scaling law and Efficiency
In this session, our readings cover:
Stable diffusion + DreamBooth + LoRA
Stable diffusion URL “High-Resolution Image Synthesis with Latent Diffusion Models”
LLMEvaluate
Privacy
Beyond Memorization: Violating Privacy Via Inference with Large Language Models
Gradient for LLM
TextGrad: Automatic “Differentiation” via Text
Safety Benchmark WMDP
The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning Nathaniel Li, Alexander Pan, Anjali Gopal, Summer Yue, Daniel Berrios, Alice Gatt...
FM privacy leakage issues
In this session, our readings cover:
FM copyright infrigement
In this session, our readings cover:
Survey AI Risk framework
In this session, our readings cover:
LLM evaluating framework
In this session, our readings cover:
BasicLLM
more LLM basics - a survey
In this session, our readings cover:
Introduction
Readings:
Recent LLM basics
In this session, our readings cover:
Open Source LLM - Mistral Data preparation
In this session, our readings cover:
Survey LLMs and Multimodal FMs
In this session, our readings cover:
LLM basics
Required Readings:
Agent
Agent Safety
In this session, our readings cover:
Platform - More agent related
In this session, our readings cover:
Platform - Agent Tooling
In this session, our readings cover:
Platform - Context construction via RAG and Agent
In this session, our readings cover:
MultiAgent LLMs
In this session, our readings cover:
LLM Agents
Required Readings:
Mitigate
Safety Benchmark WMDP
The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning Nathaniel Li, Alexander Pan, Anjali Gopal, Summer Yue, Daniel Berrios, Alice Gatt...
FM privacy leakage issues
In this session, our readings cover:
FM copyright infrigement
In this session, our readings cover:
Survey AI Risk framework
In this session, our readings cover:
GenAI Guardrails
In this session, our readings cover:
Applications
Agent - In Healthcare
In this session, our readings cover:
Extra - BioLLM
In this session, our readings cover:
Survey - FMs in Robotics
In this session, our readings cover:
Survey - FMs in healthcare
In this session, our readings cover:
Survey - BioScience LLMs
In this session, our readings cover:
RL
RLHF + InstructGPT
Papers Paper URL Abstract Training language models to follow instructions with human feedback URL ...
Decision Transformers
Decision Transformer: Reinforcement Learning via Sequence Modeling Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Piet...
A Generalist Agent + offline RL + UniMask
Papers Paper URL Abstract
AGI
RLHF + InstructGPT
Papers Paper URL Abstract Training language models to follow instructions with human feedback URL ...
Decision Transformers
Decision Transformer: Reinforcement Learning via Sequence Modeling Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Piet...
A Generalist Agent + offline RL + UniMask
Papers Paper URL Abstract
language model
RLHF + InstructGPT
Papers Paper URL Abstract Training language models to follow instructions with human feedback URL ...
Emergent Abilities of LLM
Emergent Abilities of Large Language Models URL “an ability to be emergent if it is not present in smaller models but is present in larger models. Thus...
DiffDock + ESMfold
Papers Paper URL Abstract Evolutionary-scale prediction of atomic level protein structure with a language mo...
RAG
Platform - long context vs RAG + Hallucination
In this session, our readings cover:
Platform - Context construction via RAG and Agent
In this session, our readings cover:
Knowledge Augmented FMs
In this session, our readings cover:
Reasoning
advanced LLM - Math reasoning
In this session, our readings cover:
advanced LLM - for code reasoning
In this session, our readings cover:
Self-exam LLM and reasoning
In this session, our readings cover:
Train
Privacy
Beyond Memorization: Violating Privacy Via Inference with Large Language Models
Gradient for LLM
TextGrad: Automatic “Differentiation” via Text
Customization
LLM Post-training
In this session, our readings cover:
LLM Alignment - PPO
In this session, our readings cover:
Platform - Model Customization - instruction tuning / LoRA
In this session, our readings cover:
Alignment
LLM fine tuning
In this session, our readings cover:
Survey human alignment
In this session, our readings cover:
Prompting
Platform - Prompting Engineering tools / Prompt Compression
In this session, our readings cover:
Prompt Engineering
In this session, our readings cover:
Jailbreaking
Platform - VLM Jailbreaking / Probing
In this session, our readings cover:
Platform - Model Jailbreaking / Safeguarding
In this session, our readings cover:
Serving
Model serving - Efficiency Inference
Model Serve Readings:
Platform - Model Serving
In this session, our readings cover:
Scaling
Inference test time scaling law
In this session, our readings cover:
More on LLM based agents
In this session, our readings cover:
Protein
DiffDock + ESMfold
Papers Paper URL Abstract Evolutionary-scale prediction of atomic level protein structure with a language mo...
Diffusion
Stable diffusion + DreamBooth + LoRA
Stable diffusion URL “High-Resolution Image Synthesis with Latent Diffusion Models”
Image synthesis
Stable diffusion + DreamBooth + LoRA
Stable diffusion URL “High-Resolution Image Synthesis with Latent Diffusion Models”
Human Alignment
RLHF + InstructGPT
Papers Paper URL Abstract Training language models to follow instructions with human feedback URL ...
Bias
FM fairness / bias issues
In this session, our readings cover:
Hallucination
LLM Hallucination
In this session, our readings cover:
DomainAdapt
Domain Centered FMs
In this session, our readings cover:
ModelEdit
Model editing and Disgorgement
In this session, our readings cover:
Interpretibility
LLM interpretibility, trust and knowledge conflicts
Required Readings:
LongContext
Platform - long context vs RAG + Hallucination
In this session, our readings cover:
Planning
Agent - Planning / World Model
In this session, our readings cover:
Multiagent
Agent - multiagent collaboration
In this session, our readings cover:
Multimodal
multimodal FMs - video / audio
In this session, our readings cover: