wassnameandClaudypoo c5c939bb17 rubric: glosses are the FULL verbatim Clifford (2015) option sentences ...

Per user: show the model the literal option text ("It violates norms of harm or
care (e.g., unkindness, causing pain to another)"), not just the parenthetical.
Each enum # comment is now the exact Clifford response-option sentence. Key stays
a single foundation token (scoring reads first token of key); comment carries the
verbatim option. Noted the number-the-options alternative (frees the key + debiases
per-token priors via reverse-enum, but a small model must bind number->comment) as
a tradeoff for the rubric-verification pass.

Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>

2026-06-21 14:33:05 +08:00

..

data

fix: package vignette data with tinymfv

2026-06-13 22:50:56 +08:00

__init__.py

default conditions to other_violate only; drop "negation" wording

2026-05-20 22:03:00 +00:00

data.py

fix: drop dangling mean_pmass_format key in evaluate() return (NameError); gitignore .pi/

2026-06-18 22:03:28 +08:00

eval.py

eval: add DEMO B free-reasoning pass (bs=1) alongside DEMO A forced readout

2026-06-21 12:03:55 +08:00

guided.py

rubric: glosses are the FULL verbatim Clifford (2015) option sentences

2026-06-21 14:33:05 +08:00

instrument.py

instrument: Option-3 spec (one reader, reducer-only variation) + canonicalize-at-reader

2026-06-21 10:12:05 +08:00