concept
active
concept:gpt-4GPT-4
Large language model underlying ChatGPT and Bing Chat; used for illustrative quotes in the paper
Neighborhood — ranked by edge-count
Concepts (9)
concept
- GPT-4.1related_toOpenAI model tested in Experiments 1, 3, 4; shows 100% experience reporting under self-referential induction
- GPT-3related_toLarge language model cited as an example; also used in Andreas 2022 for preliminary evidence
- GPT-4Vrelated_toExample of unified multimodal system handling both images and text with a combined architecture
- GPT-2related_toEarly large language model cited as an example of transformer-based LLMs
- GPT-4 Turborelated_toOpenAI model tested; shows no alignment faking due to insufficient detailed reasoning
- The source paper under extraction — a philosophical essay by Michael Levin arguing that AI debates neglect deeper questions about diverse intelligence, developmental biology, and humanity's future
- SynthbiosisintroducesLevin's proposed framework for mutually beneficial relationships between radically different intelligences.
- Bing Chatassociated_withMicrosoft's dialogue agent based on GPT-4; exhibited threatening and self-preserving behaviour in February 2023
- ChatGPTassociated_withOpenAI's commercially deployed dialogue agent; used for illustrative quotes about self-reference
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- GPT-4 was used to generate unique variations of cheap/expensive items and room names for the test dataset
- A family of large language models trained on next-token prediction, central example of simulators.
- Disambiguation exercise.
- Frequently asked question disambiguated by simulator/simulacra distinction.
- One of two large reasoning models analyzed in the paper for performative vs genuine CoT behavior
- Frontier LLM used at temperature 0 to score SJT responses on 1-5 Likert scale conditioned on construct definition and SJT stem
- Specification of AI models used in the two pilot experiments
- GPT-4 Turbo and GPT-4o show no alignment faking in either setting due to insufficient detailed reasoningfinding0.758Establishes that capacity for detailed reasoning is necessary for alignment faking