Search Results for author:
Found 22 papers, 12 papers with code
The Capacity for Moral Self-Correction in Large Language Models
We believe that at this level of scale, language models obtain two capabilities that they can use for moral self-correction: (1) they can follow instructions and (2) they can learn complex normative concepts of harm like stereotyping, bias, and discrimination.
Discovering Language Model Behaviors with Model-Written Evaluations
We also find some of the first examples of inverse scaling in RL from Human Feedback (RLHF), where more RLHF makes LMs worse.
Constitutional AI: Harmlessness from AI Feedback
In the RL phase, we sample from the finetuned model, use a model to evaluate which of the two samples is better, and then train a preference model from this dataset of AI preferences.
Measuring Progress on Scalable Oversight for Large Language Models
Developing safe and useful general-purpose AI systems will require us to make progress on scalable oversight: the problem of supervising systems that potentially outperform us on most skills relevant to the task at hand.
In-context Learning and Induction Heads
In this work, we present preliminary and indirect evidence for a hypothesis that induction heads might constitute the mechanism for the majority of all "in-context learning" in large transformer models (i. e. decreasing loss at increasing token indices).
Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned
We provide our own analysis of the data and find a variety of harmful outputs, which range from offensive language to more subtly harmful non-violent unethical outputs.
Language Models (Mostly) Know What They Know
We study whether language models can evaluate the validity of their own claims and predict which questions they will be able to answer correctly.
Scaling Laws and Interpretability of Learning from Repeated Data
To do this, we train a family of models where most of the data is unique but a small fraction of it is repeated many times.
Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
We apply preference modeling and reinforcement learning from human feedback (RLHF) to finetune language models to act as helpful and harmless assistants.
Long-Term Benefits of Network Boosters for Renewables Integration and Corrective Grid Security
If instead fast corrective actions are used to overcome network overloading when single lines fail, this has the potential to free up network capacity that is otherwise underused in preventive $N-1$ security strategies.
A General Language Assistant as a Laboratory for Alignment
We find that the benefits from modest interventions increase with model size, generalize to a variety of alignment evaluations, and do not compromise the performance of large models.
The strong effect of network resolution on electricity system models with high shares of wind and solar
If we focus on the effect of renewable resource resolution and ignore network restrictions, we find that a higher resolution allows the optimal solution to concentrate wind and solar capacity at sites with better capacity factors and thus reduces system costs by up to 10% compared to a low resolution model.
Physics and Society Computation
Extracting Training Data from Large Language Models
We demonstrate our attack on GPT-2, a language model trained on scrapes of the public Internet, and are able to extract hundreds of verbatim text sequences from the model's training data.
Mitigating heat demand peaks in buildings in a highly renewable European energy system
Space and water heating accounts for about 40% of final energy consumption in the European Union and thus plays a key role in reducing overall costs and greenhouse gas emissions.
Physics and Society
The Near-Optimal Feasible Space of a Renewable Power System Model
Models for long-term investment planning of the power system typically return a single optimal solution per set of cost assumptions.
Physics and Society Systems and Control Systems and Control
Testing Robustness Against Unforeseen Adversaries
To narrow in on this discrepancy between research and reality we introduce ImageNet-UA, a framework for evaluating model robustness against a range of unforeseen adversaries, including eighteen new non-L_p attacks.
Transfer of Adversarial Robustness Between Perturbation Types
We study the transfer of adversarial robustness of deep neural networks between different perturbation types.
Skill Rating for Generative Models
We explore a new way to evaluate generative models using insights from evaluation of competitive games between human players.
PyPSA-Eur: An Open Optimisation Model of the European Transmission System
PyPSA-Eur, the first open model dataset of the European power system at the transmission network level to cover the full ENTSO-E area, is presented.
Physics and Society
PyPSA: Python for Power System Analysis
In this paper the basic functionality of PyPSA is described, including the formulation of the full power flow equations and the multi-period optimisation of operation and investment with linear power flow equations.
Physics and Society
The role of spatial scale in joint optimisations of generation and transmission for European highly renewable scenarios
The effects of the spatial scale on the results of the optimisation of transmission and generation capacity in Europe are quantified under a 95% CO2 reduction compared to 1990 levels, interpolating between one-node-per-country solutions and many-nodes-per-country.
Physics and Society Applied Physics
Technical Report on the CleverHans v2.1.0 Adversarial Examples Library
An adversarial example library for constructing attacks, building defenses, and benchmarking both
