Reynolds and McDonell 2021: Multiversal views on language models

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Language models are few-shot learners (Brown et al., 2020)concept0.772
  Demonstrated transformers on mathematical understanding and logic; cited to motivate transformer versatility.
• Perez et al. 2022: Discovering language model behaviors with model-written evaluationsconcept0.767
  Study showing RLHF can exacerbate self-preservation tendencies in LLMs; key empirical support for a paper claim
• Zhu et al. 2024 - Language models represent beliefs of self and othersconcept0.764
  Key prior finding that LLMs can internally represent beliefs of self and others, motivating SOO approach
• Large language models develop surprisingly coherent yet often rigid internal preferences as they scalefinding0.763
  Mazeika et al. finding reinforcing the need for emptiness-based flexible value architectures
• Bias in language modelsconcept0.762
  Features related to gender, racial, ethnic biases, slurs, and hate speech.
• Can language models genuinely introspect on internal states or only confabulate?question0.761
  Central research question animating the paper: distinguishing genuine introspection from illusion through causal manipulation of activations.
• Towards Monosemanticity: Decomposing Language Models with Dictionary Learning (Bricken et al., 2023)concept0.761
  Foundational SAE mechanistic interpretability paper
• Can Large Language Models Genuinely Shift Human Perspectivequestion0.759