llm-control-operates-at-training-and-inference-layers

IN derived (depth 3)

Created 2026-06-21T09:57:37+00:00 · Reviewed 2026-06-21T14:41:08+00:00

LLM behavior control has developed along two axes: training-time alignment (which diversified from RLHF into three independent paradigms — full RLHF, DPO/IPO/KTO, and Constitutional AI) and inference-time prompting (which evolved from linear CoT through self-consistency to branching ToT, adding search structure to manage prompt-dependent variability). The antecedents establish these as parallel developments but do not directly establish that they compensate for each other's specific limitations.

Justifications

SL — training-time and inference-time control evolved independently into complementary stacks

Antecedents (all must be IN):

  • IN alignment-diversified-into-three-independent-paradigms — LLM alignment diversified from a single RLHF pipeline into three independent paradigms — full mathematical RLHF, direct preference optimization (DPO/IPO/KTO), and Constitutional AI — each eliminating different sources of complexity while preserving alignment quality.
  • IN prompting-sophistication-compensates-for-irreducible-sensitivity — Prompt sensitivity is a persistent, intrinsic property not resolved by scaling, and the field developed increasingly structured prompting approaches (CoT → self-consistency → tree-of-thoughts) that add search structure to reasoning. These techniques manage prompt-dependent variability by structuring the reasoning process, though the antecedents do not establish that this was the explicit motivation for their development.

Dependents

These beliefs depend on this one: