Skip to content

Reference Baseline

Training

Reference baseline. Training is included to contrast the properties of post-training techniques. It is not itself a post-training technique.

In brief — site editorial, not from the paper

The original build of a model — learning its parameters from scratch on a large dataset — included here as a reference point rather than as a post-training technique.

Nearest profiles

FT (full) (0.20), Adversarial Training (0.23), Retraining (0.33), DP-FT (0.37), MTL (0.37)

Computed from the taxonomy data (Gower distance over all six dimensions)

Cite this row

Training — six-dimensional profile
D1: parametric-update
D2: build-capability
D3: large-labeled
D4: ad-hoc-permanent, scheduled-permanent
D5: whole-model
D6: dl, fm, llm, ml, mllm
Source: arXiv:2608.06246
Compare with another technique Find in the explorer