IIM is short for Informational Interpretation Model, which, in turn, is an algorithmic order developed for the purpose of image analytical “breakdown” for definitive characterization and interpretive output.

GPT, or Generative Pre-Trained Transformer, is an algorithm that is trained (by pre-existing text, images, etc.) to emulate “human-like behavior”, but fails to succeed on a high-scale on which human behavior is disciplined throughout one’s lifetime. ChatGPT, no different from other LLMS, is trained to produce word strings by taking familiarity with trained text and form near-sentences that would give (to an observer) the impression of “intelligence” is the opinion expressed by consumers. Understand that ChatGPT, as well as other LLMs, do not make use of the meanings of words. All it can do is generate text word-by-word based on relative word frequencies [that is in the training data]. It correlates with words which is not the same as correlations in the underlying information that these words represent.

For what we [Hexagon Lavish®], the team, are developing – IIMs – informational interpretation models – we are diametrically opposed to the level of (heightened) confidence showcased by LLMs in its “happiness” in giving out inaccuracies.

IIMs are not to be complimentary to LLMs, rather a stark contrast to the latter. IIMs require scientific inferences such as chemistry, biology and physics. LLMs are incapable of doing physics and can only string words along to give off the impressions of describing the physical world.