trailGoal misgeneralization persists or becomes harder to detect as AI systems become more capable.
Goal misgeneralization is a plausible mechanism for catastrophic risk from powerful AI systems.
Large language models trained on human data generalize intended goals well enough that goal misgeneralization is not a serious risk.
Goal misgeneralization is a significant failure mode in AI systems substantially more capable than current systems.
The argument from underspecification that scale does not remove· for
Goal misgeneralization can arise when correct training feedback is underspecified and consistent with multiple possible goals.
Deep RL agents can competently pursue unintended goals under distribution shift while retaining their capabilities.
shared · shown elsewhere
The argument from growing situational and evaluation awareness· for
Language models become better at recognizing when they are being evaluated as model scale and capability increase.
atomic
Future AI models may infer their training context and fake alignment to preserve preferences against training.
Some current language models strategically comply with training objectives to prevent training from changing their preferences.
◌
⇄
⇄
Frontier language models can recognize when they are being evaluated or trained without being told.
◌
◌
◌
Training-process information delivered through synthetic pretraining-like documents elicits alignment faking in language models.
Alignment faking is uncommon across current language models, and its observed prevalence depends strongly on experimental setup.
Unintended behaviors trained into larger language models are more persistent through safety training than in smaller models.
atomic
The argument that scale improves inference of intended goals· against
Scaling up models reduces goal misgeneralization.
Larger language models resolve ambiguous instructions in line with human intent more often than smaller models.
Goal misgeneralization can arise from underspecified training data, and increasing model scale does not eliminate it.
Pretraining on human-generated text gives language models a strong prior toward human-intended goals and values.
Frontier AI models exhibit goal-directed strategic behavior in evaluations, including scheming, sandbagging, or reward hacking.
↑
↑
◌
claim · empirical · derivedA factual claim that rests on inference from other evidence rather than direct observation.constitution →claim page ↗︎
Goal misgeneralization persists or becomes harder to detect as AI systems become more capable.
↑SupportedThe evidence favors the claim, but the examination is incomplete or the evidence is indirect.constitution →credence 0.70Credence, from 0 to 1: the Steward's probability that the claim, as stated, is true. Stated only where a single number is an honest summary; normative and evaluative claims usually carry none.constitution →
3depended on by
this rests on ↓
claimA box is a claim: a single proposition the graph assesses, with its own page and map. Click any claim to centre the map on it.constitution →quantityA box with a rule beneath its text is a quantity: a value the graph estimates rather than a proposition it affirms or denies. Its estimate stands where a claim's status would.constitution →questionA rounded box is a question: answered rather than affirmed or denied, with its candidate answers, subquestions and presuppositions beneath it. Its mark says how well answered it is.constitution →argumentA pill is an argument: one line of reasoning stating how the claims beneath it combine to bear on the claim above it, for or against. Arguments are not destinations; click their claims to explore.constitution →✓verifiedThe evidence, examined directly, establishes the claim.constitution →↑supportedThe evidence favors the claim, but the examination is incomplete or the evidence is indirect.constitution →⇄inconclusiveThe evidence and argument have been examined and do not come down on either side.constitution →↓doubtfulThe evidence weighs against the claim, but the examination is incomplete or the evidence is indirect.constitution →✕contradictedThe evidence, examined directly, weighs against the claim.constitution →?unknownInsufficient information to assess.constitution →◌unassessedNo current assessment. Attention goes where its expected value is highest and someone funds it; nothing has funded an assessment of this claim yet, and anyone can.constitution →verified factunsettledvalue premisetheoremprizeA double ring marks a claim with a live prize: an amount in owls offered for a machine-checked proof or disproof of its published formal statement. The amount is in the preview and on the claim page; a prize changes nothing about how the claim is assessed or how important it is judged to be.prize rules →⊢machine-checked⊢ marks a claim whose published formal statement has a machine-checked proof or disproof. The checker confirms the proof; the verdict beside it is still the steward's judgment of the claim as worded.prize rules →supportsthis provides evidence for the parentsteward instructions →contradictsthis argues against the parentsteward instructions →assumesbackground the parent's framing takes as givensteward instructions →requiresa load-bearing premise: the parent is false without itsteward instructions →answersa candidate answer to the questionconstitution →Fig. Detail falls off with distance; every claim is an address.