CoChat

Me, Myself, and AI: The Situational Awareness Dataset (SAD) for LLMs

Open access
Laine, Rudolf, Chughtai, Bilal, Betley, Jan, Hariharan, Kaivalya, Scheurer, Jérémy, Balesni, Mikita, Hobbhahn, Marius, Meinke, Alexander, Evans, Owain · 2024 · Advances in Neural Information Processing Systems 37

The DOI matches its Crossref record; 8 of the 8 references it cites were found in a registry or an index.

Abstract

AI assistants such as ChatGPT are trained to respond to users by saying, "I am a large language model". This raises questions. Do such models know that they are LLMs and reliably act on this knowledge? Are they aware of their current circumstances, such as being deployed to the public? We refer to a model's knowledge of itself and its circumstances as situational awareness. To quantify situational awareness in LLMs, we introduce a range of behavioral tests, based on question answering and instruction following. These tests form the $\textbf{Situational Awareness Dataset (SAD)}$, a benchmark comprising 7 task categories and over 13,000 questions. The benchmark tests numerous abilities, including the capacity of LLMs to (i) recognize their own generated text, (ii) predict their own behavior, (iii) determine whether a prompt is from internal evaluation or real-world deployment, and (iv) follow instructions that depend on self-knowledge. We evaluate 16 LLMs on SAD, including both base (pretrained) and chat models. While all models perform better than chance, even the highest-scoring model (Claude 3 Opus) is far from a human baseline on certain tasks. We also observe that performance on SAD is only partially predicted by metrics of general knowledge (e.g. MMLU). Chat models, which are finetuned to serve as AI assistants, outperform their corresponding base models on SAD but not on general knowledge tasks. The purpose of SAD is to facilitate scientific understanding of situational awareness in LLMs by breaking it down into quantitative abilities. Situational awareness is important because it enhances a model's capacity for autonomous planning and action. While this has potential benefits for automation, it also introduces novel risks related to AI safety and control. Code and latest results available at https://situational-awareness-dataset.org .

Overview

Integrity signals

Registry

Crossref · 10.52202/079017-2044

Cites

8 of 8 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Verification

Claims the paper makes

We couldn’t retrieve this paper’s full text, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Show all 8 references

Cited by

See all 8

Metadata

Authors
Laine, R., Chughtai, B., Betley, J., Hariharan, K., Scheurer, J., Balesni, M., Hobbhahn, M., Meinke, A., & Evans, O.
Venue
Advances in Neural Information Processing Systems 37
Published
2024
Pages
64010-64118
ISSN
1049-5258 (ISSN-L)
OpenAlex
W4415797398
Cited by
8
References
8
Citations by year
2026: 6 · 2025: 2
FWCI
2.31
Type
conference paper
Language
en
Open access
Licence
public-domain
Locations
  • neural information processing systems (publishedVersion)
  • arXiv (Cornell University) (submittedVersion, public-domain)
  • arXiv (Cornell University)
Funders
Open Philanthropy Project
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
  • Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
  • Social Sciences › Human Factors and Ergonomics › Human-Automation Interaction and Safety
Keywords
situational awareness, large language models, human-AI interaction, datasets, human-AI collaboration

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Me, Myself, and AI: The Situational Awareness Dataset (SAD) for LLMs. https://app.cochat.ai/papers/10.52202/079017-2044
The paper
Laine, R., Chughtai, B., Betley, J., Hariharan, K., Scheurer, J., Balesni, M., Hobbhahn, M., Meinke, A., & Evans, O. (2024). Me, Myself, and AI: The Situational Awareness Dataset (SAD) for LLMs. Advances in Neural Information Processing Systems 37. https://doi.org/10.52202/079017-2044

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.