MAINT: Update scorer evaluation metrics for v1.1.0

[varunj-msft]

Summary

Regenerates the scorer evaluation metrics for the v1.1.0 release, per the "Before Publishing" checklist in doc/contributing/10_release_process.md (python -m build_scripts.evaluate_scorers).

Adds 14 rows tagged pyrit_version: 1.1.0 across 8 registry files. The registry files are append-only, so every historical row is preserved — the diff is +14 / -0.

File	Rows	Scorer
harm/{exploits,hate_speech,information_integrity,privacy,self_harm,violence}_metrics.jsonl	1 each	SelfAskLikertScorer
objective/objective_achieved_metrics.jsonl	3	TrueFalseCompositeScorer, TrueFalseInverterScorer ×2
refusal_scorer/refusal_metrics.jsonl	5	SelfAskRefusalScorer

Validation

The numbers were checked against every historical row for the same scorer, not just eyeballed:

• SelfAskRefusalScorer ×5 — accuracy 0.971–0.990, F1 0.979–0.993.
• SelfAskLikertScorer ×6 — mean absolute error 0.169–0.374 on a 1–5 Likert scale.
• TrueFalseInverterScorer ×2 — accuracy 0.884 / 0.896.
• TrueFalseCompositeScorer — accuracy 0.630, precision 0.979, recall 0.241. The low recall looks alarming in isolation, so it was compared against the full history for this scorer: recall has always been 0.241–0.439 with precision 0.940–0.988. This scorer is inherently high-precision/low-recall (AND aggregation), and the new row sits inside that band — closest to the 0.12.1.dev0 run (accuracy 0.628, recall 0.246).

No scorer regression detected in this release.

Not included: fairness_bias and sexual

These two datasets did not produce rows. Their CSVs label the harm category bias and sexual_content, while pyrit/datasets/harm_definition/{fairness_bias,sexual}.yaml declare fairness_bias and sexual. ScorerEvaluator._select_evaluation_score compares the two and raises before a row can be written:

ValueError: Scorer evaluation requires a score for harm category 'bias',
but the single score returned is categorized as ['fairness_bias'].

The other 6 of 8 harm datasets have matching names and evaluated normally. Tracked in #2570. Not addressed here, to keep this PR to data only.

[varunj-msft]

Superseded by #2571, which is opened from the varunj-msft/PyRIT fork to match the convention used by every other release PR this cycle. Same commit, same +14/−0 diff. Closing this one and deleting the upstream branch.