Model Comparison
Model Editorial Structural Class Conf SETL Theme
@cf/meta/llama-3.3-70b-instruct-fp8-fast lite +0.40 ND Moderate positive 0.80 0.00 Media Integrity
@cf/meta/llama-4-scout-17b-16e-instruct lite +0.10 ND Mild positive 0.80 0.00 Media Integrity
Section @cf/meta/llama-3.3-70b-instruct-fp8-fast lite @cf/meta/llama-4-scout-17b-16e-instruct lite Delta
Preamble ND ND
Article 1 ND ND
Article 2 ND ND
Article 3 ND ND
Article 4 ND ND
Article 5 ND ND
Article 6 ND ND
Article 7 ND ND
Article 8 ND ND
Article 9 ND ND
Article 10 ND ND
Article 11 ND ND
Article 12 ND ND
Article 13 ND ND
Article 14 ND ND
Article 15 ND ND
Article 16 ND ND
Article 17 ND ND
Article 18 ND ND
Article 19 ND ND
Article 20 ND ND
Article 21 ND ND
Article 22 ND ND
Article 23 ND ND
Article 24 ND ND
Article 25 ND ND
Article 26 ND ND
Article 27 ND ND
Article 28 ND ND
Article 29 ND ND
Article 30 ND ND
+0.40 A writer fabricated a series of stories for Atlas Obscura (thewalrus.ca)
220 points by pseudolus 761 days ago | 155 comments on HN | Moderate positive ~lite vlite-1.4
Summary ~lite Media Integrity Acknowledges
Exposes fabrication
EQ 0.60
SO 0.40
TD 0.70
Lite evaluation by llama-3.3-70b-wai · editorial channel only · no per-section breakdown available
Audit Trail 11 entries
2026-02-28 10:14 model_divergence Cross-model spread 0.30 exceeds threshold (2 models) - -
2026-02-28 10:14 eval_success Lite evaluated: Mild positive (0.10) - -
2026-02-28 10:14 eval Evaluated by llama-4-scout-wai: +0.10 (Mild positive) 0.00
2026-02-28 10:14 rater_validation_warn Lite validation warnings for model llama-4-scout-wai: 0W 1R - -
2026-02-28 10:09 model_divergence Cross-model spread 0.30 exceeds threshold (2 models) - -
2026-02-28 10:09 eval_success Lite evaluated: Mild positive (0.10) - -
2026-02-28 10:09 eval Evaluated by llama-4-scout-wai: +0.10 (Mild positive)
2026-02-28 10:09 rater_validation_warn Lite validation warnings for model llama-4-scout-wai: 0W 1R - -
2026-02-28 09:55 eval_success Lite evaluated: Moderate positive (0.40) - -
2026-02-28 09:55 rater_validation_warn Lite validation warnings for model llama-3.3-70b-wai: 0W 1R - -
2026-02-28 09:55 eval Evaluated by llama-3.3-70b-wai: +0.40 (Moderate positive)