Alignment Newsletter Podcast Rohin Shah et al.

Recorded by Robert Miles: http://robertskmiles.com
More information about the newsletter here: https://rohinshah.com/alignment-newsletter/
YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
  HIGHLIGHTS Scaling Language Models: Methods, Analysis & Insights from Training Gopher (Jack W. Rae et al) (summarized by Rohin): This paper details the training of the Gopher family of large language models (LLMs), the biggest of which is named Gopher and has 280 billion parameters. The algorithmic details are very similar to the GPT series (AN #102): a Transformer architecture trained on next-word prediction. The models are trained on a new data distribution that still consists of text from the Internet but in different proportions (for example, book data is 27% of Gopher’s training data but only 16% of GPT-3’s training data).
Like other LLM papers, there are tons of evaluations of Gopher on various tasks, only some of which I’m going to cover here. One headline number is that Gopher beat the state of the art (SOTA) at the time on 100 out of 124 evaluation tasks.
The most interesting aspect of the paper (to me) is that the entire Gopher family of models were all trained on the same number of tokens, thus allowing us to study the effect of scaling up model parameters (and thus training compute) while holding data constant. Some of the largest benefits of scale were seen in the Medicine, Science, Technology, Social Sciences, and the Humanities task categories, while scale has not much effect or even a negative effect in the Maths, Logical Reasoning, and Common Sense categories. Surprisingly, we see improved performance on TruthfulQA (AN #165) with scale, even though the TruthfulQA benchmark was designed to show worse performance with increased scale.
We can use Gopher in a dialogue setting by prompting it appropriately. The prompt specifically instructs Gopher to be “respectful, polite, and inclusive”; it turns out that this significantly helps with toxicity. In particular, for the vanilla Gopher model family, with more scale the models produce more toxic continuations given toxic user statements; this no longer happens with Dialogue-Prompted Gopher models, which show slight reductions in toxicity with scale in the same setting. The authors speculate that while increased scale leads to an increased ability to mimic the style of a user statement, this is compensated for by an increased ability to account for the prompt.
Another alternative the authors explore is to finetune Gopher on 5 billion tokens of dialogue to produce Dialogue-Tuned Gopher. Interestingly, human raters were indifferent between Dialogue-Prompted Gopher and Dialogue-Tuned Gopher.
Read more: Blog post: Language modelling at scale: Gopher, ethical considerations, and retrieval
Training Compute-Optimal Large Language Models (Jordan Hoffmann et al) (summarized by Rohin): One application of scaling laws (AN #87) is to figure out how big a model to train, on how much data, given some compute budget. This paper performs a more systematic study than the original paper and finds that existing models are significantly overtrained. Chinchilla is a new model built with this insight: it has 4x fewer parameters than Gopher, but is trained on 4x as much data. Despite using the same amount of training compute as Gopher (and lower inference compute), Chinchilla outperforms Gopher across a wide variety of metrics, validating these new scaling laws.
You can safely skip to the opinion at this point – the rest of this summary is quantitative details.
We want to find functions N(C) and D(C) that specify the optimal number of parameters N and the amount of data D to use given some compute budget C. We’ll assume that these scale with a power of C, that is, N(C) = k_N * C^a and D(C) = k_D * C^b, for some constants a, b, k_N, and k_D. Note that since total compute increases linearly with both N (since each forwa

Recorded by Robert Miles: http://robertskmiles.com
More information about the newsletter here: https://rohinshah.com/alignment-newsletter/
YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
 
Sorry for the long hiatus! I was really busy over the past few months and just didn't find time to write this newsletter. (Realistically, I was also a bit tired of writing it and so lacked motivation.) I'm intending to go back to writing it now, though I don't think I can realistically commit to publishing weekly; we'll see how often I end up publishing. For now, have a list of all the things I should have advertised to you whose deadlines haven't already passed.   NEWS Survey on AI alignment resources (Anonymous) (summarized by Rohin): This survey is being run by an outside collaborator in partnership with the Centre for Effective Altruism (CEA). They ask that you fill it out to help field builders find out which resources you have found most useful for learning about and/or keeping track of the AI alignment field. Results will help inform which resources to promote in the future, and what type of resources we should make more of.
Announcing the Inverse Scaling Prize ($250k Prize Pool) (Ethan Perez et al) (summarized by Rohin): This prize with a $250k prize pool asks participants to find new examples of tasks where pretrained language models exhibit inverse scaling: that is, models get worse at the task as they are scaled up. Notably, you do not need to know how to program to participate: a submission consists solely of a dataset giving at least 300 examples of the task.
Inverse scaling is particularly relevant to AI alignment, for two main reasons. First, it directly helps understand how the language modeling objective ("predict the next word") is outer misaligned, as we are finding tasks where models that do better according to the language modeling objective do worse on the task of interest. Second, the experience from examining inverse scaling tasks could lead to general observations about how best to detect misalignment.
$500 bounty for alignment contest ideas (Akash) (summarized by Rohin): The authors are offering a $500 bounty for producing a frame of the alignment problem that is accessible to smart high schoolers/college students and people without ML backgrounds. (See the post for details; this summary doesn't capture everything well.)
Job ad: Bowman Group Open Research Positions (Sam Bowman) (summarized by Rohin): Sam Bowman is looking for people to join a research center at NYU that'll focus on empirical alignment work, primarily on large language models. There are a variety of roles to apply for (depending primarily on how much research experience you already have).
Job ad: Postdoc at the Algorithmic Alignment Group (summarized by Rohin): This position at Dylan Hadfield-Menell's lab will lead the design and implementation of a large-scale Cooperative AI contest to take place next year, alongside collaborators at DeepMind and the Cooperative AI Foundation.
Job ad: AI Alignment postdoc (summarized by Rohin): David Krueger is hiring for a postdoc in AI alignment (and is also hiring for another role in deep learning). The application deadline is August 2.
Job ad: OpenAI Trust & Safety Operations Contractor (summarized by Rohin): In this remote contractor role, you would evaluate submissions to OpenAI's App Review process to ensure they comply with OpenAI's policies. Apply here by July 13, 5pm Pacific Time.
Job ad: Director of CSER (summarized by Rohin): Application deadline is July 31. Quoting the job ad: "The Director will be expected to provide visionary leadership for the Centre, to maintain and enhance its reputation for cutting-edge research, to develop and oversee fundraising and new project and programme design, to ensure the proper functioning of its operations and administration, and to lead its endeavours to secure

Recorded by Robert Miles: http://robertskmiles.com
More information about the newsletter here: https://rohinshah.com/alignment-newsletter/
YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
 
HIGHLIGHTS Alignment difficulty (Richard Ngo and Eliezer Yudkowsky) (summarized by Rohin): Eliezer is known for being pessimistic about our chances of averting AI catastrophe. His argument in this dialogue is roughly as follows:
1. We are very likely going to keep improving AI capabilities until we reach AGI, at which point either the world is destroyed, or we use the AI system to take some pivotal act before some careless actor destroys the world.
2. In either case, the AI system must be producing high-impact, world-rewriting plans; such plans are “consequentialist” in that the simplest way to get them (and thus, the one we will first build) is if you are forecasting what might happen, thinking about the expected consequences, considering possible obstacles, searching for routes around the obstacles, etc. If you don’t do this sort of reasoning, your plan goes off the rails very quickly - it is highly unlikely to lead to high impact. In particular, long lists of shallow heuristics (as with current deep learning systems) are unlikely to be enough to produce high-impact plans.
3. We’re producing AI systems by selecting for systems that can do impressive stuff, which will eventually produce AI systems that can accomplish high-impact plans using a general underlying “consequentialist”-style reasoning process (because that’s the only way to keep doing more impressive stuff). However, this selection process does not constrain the goals towards which those plans are aimed. In addition, most goals seem to have convergent instrumental subgoals like survival and power-seeking that would lead to extinction. This suggests that we should expect an existential catastrophe by default.
4. None of the methods people have suggested for avoiding this outcome seem like they actually avert this story.
Richard responds to this with a few distinct points:
1. It might be possible to build AI systems which are not of world-destroying intelligence and agency, that humans use to save the world. For example, we could make AI systems that do better alignment research. Such AI systems do not seem to require the property of making long-term plans in the real world in point (3) above, and so could plausibly be safe.
2. It might be possible to build general AI systems that only state plans for achieving a goal of interest that we specify, without executing that plan.
3. It seems possible to create consequentialist systems with constraints upon their reasoning that lead to reduced risk.
4. It also seems possible to create systems with the primary aim of producing plans with certain properties (that aren't just about outcomes in the world) -- think for example of corrigibility (AN #35) or deference to a human user.
5. (Richard is also more bullish on coordinating not to use powerful and/or risky AI systems, though the debate did not discuss this much.)
Eliezer’s responses:
1. AI systems that help with alignment research to such a degree that it actually makes a difference are almost certainly already dangerous.
2. It is the plan itself that is risky; if the AI system made a plan for a goal that wasn’t the one we actually meant, and we don’t understand that plan, that plan can still cause extinction. It is the misaligned optimization that produced the plan that is dangerous.
3 and 4. It is certainly possible to do such things; the space of minds that could be designed is very large. However, it is difficult to do such things, as they tend to make consequentialist reasoning weaker, and on our current trajectory the first AGI that we build will probably not look like that.
This post has also been summarized by others here, though with different emphases than in my summary.
 

Recorded by Robert Miles: http://robertskmiles.com
More information about the newsletter here: https://rohinshah.com/alignment-newsletter/
YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
  HIGHLIGHTS Draft report on existential risk from power-seeking AI (Joe Carlsmith) (summarized by Rohin): This report investigates the classic AI risk argument in detail, and decomposes it into a set of conjunctive claims. Here’s the quick version of the argument. We will likely build highly capable and agentic AI systems that are aware of their place in the world, and which will be pursuing problematic objectives. Thus, they will take actions that increase their power, which will eventually disempower humans leading to an existential catastrophe. We will try and avert this, but will probably fail to do so since it is technically challenging, and we are not capable of the necessary coordination.
There’s a lot of vague words in the argument above, so let’s introduce some terminology to make it clearer:
- Advanced capabilities: We say that a system has advanced capabilities if it outperforms the best humans on some set of important tasks (such as scientific research, business/military/political strategy, engineering, and persuasion/manipulation).
- Agentic planning: We say that a system engages in agentic planning if it (a) makes and executes plans, (b) in pursuit of objectives, (c) on the basis of models of the world. This is a very broad definition, and doesn’t have many of the connotations you might be used to for an agent. It does not need to be a literal planning algorithm -- for example, human cognition would count, despite (probably) not being just a planning algorithm.
- Strategically aware: We say that a system is strategically aware if it models the effects of gaining and maintaining power over humans and the real-world environment.
- PS-misaligned (power-seeking misaligned): On some inputs, the AI system seeks power in unintended ways, due to problems with its objectives (if the system actually receives such inputs, then it is practically PS-misaligned.)
The core argument is then that AI systems with advanced capabilities, agentic planning, and strategic awareness (APS-systems) will be practically PS-misaligned, to an extent that causes an existential catastrophe. Of course, we will try to prevent this -- why should we expect that we can’t fix the problem? The author considers possible remedies, and argues that they all seem quite hard:
- We could give AI systems the right objectives (alignment), but this seems quite hard -- it’s not clear how we would solve either outer or inner alignment.
- We could try to shape objectives to be e.g. myopic, but we don’t know how to do this, and there are strong incentives against myopia.
- We could try to limit AI capabilities by keeping systems special-purpose rather than general, but there are strong incentives for generality, and some special-purpose systems can be dangerous, too.
- We could try to prevent the AI system from improving its own capabilities, but this requires us to anticipate all the ways the AI system could improve, and there are incentives to create systems that learn and change as they gain experience.
- We could try to control the deployment situations to be within some set of circumstances where we know the AI system won’t seek power. However, this seems harder and harder to do as capabilities increase, since with more capabilities, more options become available.
- We could impose a high threshold of safety before an AI system is deployed, but the AI system could still seek power during training, and there are many incentives pushing for faster, riskier deployment (even if we have already seen warning shots).
- We could try to correct the behavior of misaligned AI systems, or mitigate their impact, after deployment. This seems like it requires humans to have comparabl

Recorded by Robert Miles: http://robertskmiles.com
More information about the newsletter here: https://rohinshah.com/alignment-newsletter/
YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
 
HIGHLIGHTS Collaborating with Humans without Human Data (DJ Strouse et al) (summarized by Rohin): We’ve previously seen that if you want to collaborate with humans in the video game Overcooked, it helps to train a deep RL agent against a human model (AN #70), so that the agent “expects” to be playing against humans (rather than e.g. copies of itself, as in self-play). We might call this a “human-aware” model. However, since a human-aware model must be trained against a model that imitates human gameplay, we need to collect human gameplay data for training. Could we instead train an agent that is robust enough to play with lots of different agents, including humans as a special case?
This paper shows that this can be done with Fictitious Co-Play (FCP), in which we train our final agent against a population of self-play agents and their past checkpoints taken throughout training. Such agents get significantly higher rewards when collaborating with humans in Overcooked (relative to the human-aware approach in the previously linked paper).
In their ablations, the authors find that it is particularly important to include past checkpoints in the population against which you train. They also test whether it helps to have the self-play agents have a variety or architectures, and find that it mostly does not make a difference (as long as you are using past checkpoints as well).
Read more: Related paper: Maximum Entropy Population Based Training for Zero-Shot Human-AI Coordination
Rohin's opinion: You could imagine two different philosophies on how to build AI systems -- the first option is to train them on the actual task of interest (for Overcooked, training agents to play against humans or human models), while the second option is to train a more robust agent on some more general task, that hopefully includes the actual task within it (the approach in this paper). Besides Overcooked, another example would be supervised learning on some natural language task (the first philosophy), as compared to pretraining on the Internet GPT-style and then prompting the model to solve your task of interest (the second philosophy). In some sense the quest for a single unified AGI system is itself a bet on the second philosophy -- first you build your AGI that can do all tasks, and then you point it at the specific task you want to do now.
Historically, I think AI has focused primarily on the first philosophy, but recent years have shown the power of the second philosophy. However, I don’t think the question is settled yet: one issue with the second philosophy is that it is often difficult to fully “aim” your system at the true task of interest, and as a result it doesn’t perform as well as it “could have”. In Overcooked, the FCP agents will not learn specific quirks of human gameplay that could be exploited to improve efficiency (which the human-aware agent could do, at least in theory). In natural language, even if you prompt GPT-3 appropriately, there’s still some chance it ends up rambling about something else entirely, or neglects to mention some information that it “knows” but that a human on the Internet would not have said. (See also this post (AN #141).)
I should note that you can also have a hybrid approach, where you start by training a large model with the second philosophy, and then you finetune it on your task of interest as in the first philosophy, gaining the benefits of both.
I’m generally interested in which approach will build more useful agents, as this seems quite relevant to forecasting the future of AI (which in turn affects lots of things including AI alignment plans).
  TECHNICAL AI ALIGNMENT

Recorded by Robert Miles: http://robertskmiles.com
More information about the newsletter here: https://rohinshah.com/alignment-newsletter/
YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
 
HIGHLIGHTS Request for proposals for projects in AI alignment that work with deep learning systems (Nick Beckstead and Asya Bergal) (summarized by Rohin): Open Philanthropy is seeking proposals for AI safety work in four major areas related to deep learning, each of which I summarize below. Proposals are due January 10, and can seek up to $1M covering up to 2 years. Grantees may later be invited to apply for larger and longer grants.
Rohin's opinion: Overall, I like these four directions and am excited to see what comes out of them! I'll comment on specific directions below.
RFP: Measuring and forecasting risks (Jacob Steinhardt) (summarized by Rohin): Measurement and forecasting is useful for two reasons. First, it gives us empirical data that can improve our understanding and spur progress. Second, it can allow us to quantitatively compare the safety performance of different systems, which could enable the creation of safety standards. So what makes for a good measurement?
1. Relevance to AI alignment: The measurement exhibits a failure mode that becomes worse as models become larger, or tracks a potential capability that may emerge with further scale (which in turn could enable deception, hacking, resource acquisition, etc).
2. Forward-looking: The measurement helps us understand future issues, not just those that exist today. Isolated examples of a phenomenon are good if we have nothing else, but we’d much prefer to have a systematic understanding of when a phenomenon occurs and how it tends to quantitatively increase or decrease with various factors. See for example scaling laws (AN #87).
3. Rich data source: Not all trends in MNIST generalize to CIFAR-10, and not all trends in CIFAR-10 generalize to ImageNet. Measurements on data sources with rich factors of variation are more likely to give general insights.
4. Soundness and quality: This is a general category for things like “do we know that the signal isn’t overwhelmed by the noise” and “are there any reasons that the measurement might produce false positives or false negatives”.
What sorts of things might you measure?
1. As you scale up task complexity, how much do you need to scale up human-labeled data to continue to maintain good performance and avoid reward hacking? If you fail at this and there are imperfections in the reward, how bad does this become?
2. What changes do we observe based on changes in the quality of the human feedback (e.g. getting feedback from amateurs vs experts)? This could give us information about the acceptable “difference in intelligence” between a model and its supervisor.
3. What happens when models are pushed out of distribution along a factor of variation that was not varied in the pretraining data?
4. To what extent do models provide wrong or undesired outputs in contexts where they are capable of providing the right answer?
Rohin's opinion: Measurements generally seem great. One story for impact is that we have a measurement that we think is strongly correlated with x-risk, and we use that measurement to select an AI system that scores low on such a metric. This seems distinctly good and I think would in fact reduce x-risk! But I want to clarify that I don’t think it would convince me that the system was safe with high confidence. The conceptual arguments against high confidence in safety seem quite strong and not easily overcome by such measurements. (I’m thinking of objective robustness failures (AN #66) of the form “the model is trying to pursue a simple proxy, but behaves well on the training distribution until it can execute a treacherous turn”.)
You can also tell stories where the