flat autoregressive LLMs

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

• Autoregressive modelsframework0.784
  Second model system studied; used to show why flat autoregressive LLMs struggle with long-range coherence.
• Autoregressive Samplingmethod0.754
  The mechanism by which LLMs generate text: drawing a token from the next-token distribution and appending it to context repeatedly
• Reflection in LLMsconcept0.749
  The core phenomenon studied: the ability of LLMs to evaluate and revise their own reasoning.
• autoregressive persistenceconcept0.746
  Baseline persistence of any probe direction arising from the autoregressive nature of LLMs, not specific to emotion content
• autoregressive recurrenceconcept0.745
  Transformers are recurrent through autoregression because the K/V stream provides horizontal information flow across positions, even though each forward pass is feedforward.
• LLM-Judge Data Attributionmethod0.743
  Alternative data attribution approach using an LLM as a judge; compared against the probe-based method.
• autoregressive modelingmethod0.740
  Statistical technique where outputs are regressed on previous values; used in language generation
• To what extent is there persistence of emotional state beyond what is expected merely from the autoregressive nature of LLMs?question0.734
  The central research question motivating the paper