wassnameandClaudypoo 40d08f8d7a Establish P4 WVS map data foundation (GlobalOpinionQA) ...

Probe + artifact: the WVS subset of Anthropic/llm_global_opinions is 353 questions
over 90 countries (212 questions with >=40 countries), matching tinymfv's MC +
human-anchor shape and dense enough for an Economist-scale map. Documents the
selections parse recipe and the open axis-definition fork (literal IW 10-question
factor model vs shared-question ipsative PCA) before model-run compute.

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

2026-07-04 19:43:24 +08:00

..

05_upload_hf.py

cleanup: remove legacy dataset creation artifacts

2026-06-25 20:21:49 +08:00

09_forced_choice.py

api: clarify eval readout names

2026-06-25 20:22:04 +08:00

bootstrap_mfq2_sample_n.py

Export MFQ2 sample logprobs for N bootstrap

2026-07-01 06:24:13 +08:00

plot_steer_showcase.py

Add MFQ-2 per-zone p90 respondent ellipses to ipsative map

2026-07-04 19:25:52 +08:00

probe_api_readout.py

Add sampling readout for logprob-less API models (OpenRouter)

2026-07-04 19:39:18 +08:00

probe_mfv_think_budget.py

probe: add n_samples/temperature for BMA top1 test

2026-06-25 03:34:34 +08:00

probe_word_readout.py

add word-readout diagnostic (confirm validation-table 0.826 under old method)

2026-06-25 04:22:14 +08:00

probe_wvs_foundation.py

Establish P4 WVS map data foundation (GlobalOpinionQA)

2026-07-04 19:43:24 +08:00

summarize_showcase_effects.py

Gate showcase paths on contrast and margin

2026-06-30 14:01:59 +08:00