research-document

FE-011A Pilot Run Log

Run 1

run_id: FE-011A-SWOT-001
date: 2026-06-28
model: GPT-5 Codex
framework: SWOT
version_label: Version B
packet_path: research/experiments/FE-011A-llm-blind-pilot/framework-packets/swot-version-b.md
scenario_path: research/experiments/FE-011A-llm-blind-pilot/scenarios/swot-scenario.md
participant_prompt_path: research/experiments/FE-011A-llm-blind-pilot/prompts/participant-prompt.md
output_path: research/experiments/FE-011A-llm-blind-pilot/results/swot/output-1.md
evaluator: GPT-5 Codex blinded evaluator pass
evaluator_output_path: research/experiments/FE-011A-llm-blind-pilot/results/swot/evaluator-review.md
notes: Output was stored as Output 1 for blinded comparison.

run_id: FE-011A-SWOT-002
date: 2026-06-28
model: GPT-5 Codex
framework: SWOT
version_label: Version A
packet_path: research/experiments/FE-011A-llm-blind-pilot/framework-packets/swot-version-a.md
scenario_path: research/experiments/FE-011A-llm-blind-pilot/scenarios/swot-scenario.md
participant_prompt_path: research/experiments/FE-011A-llm-blind-pilot/prompts/participant-prompt.md
output_path: research/experiments/FE-011A-llm-blind-pilot/results/swot/output-2.md
evaluator: GPT-5 Codex blinded evaluator pass
evaluator_output_path: research/experiments/FE-011A-llm-blind-pilot/results/swot/evaluator-review.md
notes: Output was stored as Output 2 for blinded comparison.

run_id: FE-011A-5W-001
date: 2026-06-28
model: GPT-5 Codex
framework: Five Whys
version_label: Version B
packet_path: research/experiments/FE-011A-llm-blind-pilot/framework-packets/five-whys-version-b.md
scenario_path: research/experiments/FE-011A-llm-blind-pilot/scenarios/five-whys-scenario.md
participant_prompt_path: research/experiments/FE-011A-llm-blind-pilot/prompts/participant-prompt.md
output_path: research/experiments/FE-011A-llm-blind-pilot/results/five-whys/output-1.md
evaluator: GPT-5 Codex blinded evaluator pass
evaluator_output_path: research/experiments/FE-011A-llm-blind-pilot/results/five-whys/evaluator-review.md
notes: Output was stored as Output 1 for blinded comparison.

run_id: FE-011A-5W-002
date: 2026-06-28
model: GPT-5 Codex
framework: Five Whys
version_label: Version A
packet_path: research/experiments/FE-011A-llm-blind-pilot/framework-packets/five-whys-version-a.md
scenario_path: research/experiments/FE-011A-llm-blind-pilot/scenarios/five-whys-scenario.md
participant_prompt_path: research/experiments/FE-011A-llm-blind-pilot/prompts/participant-prompt.md
output_path: research/experiments/FE-011A-llm-blind-pilot/results/five-whys/output-2.md
evaluator: GPT-5 Codex blinded evaluator pass
evaluator_output_path: research/experiments/FE-011A-llm-blind-pilot/results/five-whys/evaluator-review.md
notes: Output was stored as Output 2 for blinded comparison.

run_id: FE-011A-OODA-001
date: 2026-06-28
model: GPT-5 Codex
framework: OODA
version_label: Version A
packet_path: research/experiments/FE-011A-llm-blind-pilot/framework-packets/ooda-version-a.md
scenario_path: research/experiments/FE-011A-llm-blind-pilot/scenarios/ooda-scenario.md
participant_prompt_path: research/experiments/FE-011A-llm-blind-pilot/prompts/participant-prompt.md
output_path: research/experiments/FE-011A-llm-blind-pilot/results/ooda/output-1.md
evaluator: GPT-5 Codex blinded evaluator pass
evaluator_output_path: research/experiments/FE-011A-llm-blind-pilot/results/ooda/evaluator-review.md
notes: Output was stored as Output 1 for blinded comparison.

run_id: FE-011A-OODA-002
date: 2026-06-28
model: GPT-5 Codex
framework: OODA
version_label: Version B
packet_path: research/experiments/FE-011A-llm-blind-pilot/framework-packets/ooda-version-b.md
scenario_path: research/experiments/FE-011A-llm-blind-pilot/scenarios/ooda-scenario.md
participant_prompt_path: research/experiments/FE-011A-llm-blind-pilot/prompts/participant-prompt.md
output_path: research/experiments/FE-011A-llm-blind-pilot/results/ooda/output-2.md
evaluator: GPT-5 Codex blinded evaluator pass
evaluator_output_path: research/experiments/FE-011A-llm-blind-pilot/results/ooda/evaluator-review.md
notes: Output was stored as Output 2 for blinded comparison.

This pilot was executed as an internal single-agent run.
The same model family generated participant outputs and evaluator reviews.
Findings should therefore be treated as pilot evidence about the instrument, not as independent validation.