sent_tokenize (NLTK sentence tokenizer)

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

• tokenizer vocabularyconcept0.805
  The standard set of tokens that the functional token remains a part of.
• NLTK Stemming and Lemmatizationmethod0.721
  Used to normalize candidate instruction tokens in the instruction discovery experiment.
• Stochastic text generation (next token prediction)method0.711
  The core mechanism of LLMs: predicting the next token based on previous context.
• Digit-token logit distributionconcept0.700
  Full distribution over tokens 0-9 at first generation step; contains more information than any single sampled token
• Natural Language Autoencoders (NLA)framework0.698
  An unsupervised method for generating natural language explanations of LLM activations through a verbalizer-reconstructor pair trained jointly with RL.
• Token-in-Context Featureconcept0.696
  Feature that fires on a specific token only within a specific surrounding context (e.g., 'the' in physics vs 'the' in mathematics)
• Natural Language Autoencoders (NLAs)method0.695
  Core unsupervised method for generating natural language explanations of LLM activations through a verbalizer-reconstructor pair trained with RL.
• What class of sentences to allow as expressing various kinds of commitments?question0.691
  Open problem on the expressiveness of commitment sentences.