實驗EXP-2026-0101v0.1

Experiment A——單次智能與鷹架增益的受控實驗協定 v0.1

第一個 v0.2 實驗：對同一模型與固定任務集，沿鷹架階梯 A0 原生單次 → A1 放大軌跡 → A2 多樣本 → A3 驗證器 → A4 確定性本地工具 → A5 有界完整 agentic 迴圈往上爬，每一級同時記品質與物理成本，得出鷹架響應曲線、SSR/SDR、鷹架成本倍率與邊際產率。五個假說（H1 增益存在、H2 增益有物理成本、H3 邊際產率非常數、H4 原生與系統能力可區分、H5 不同模型有不同鷹架剖面）、預先登記的品質投影、A0–A3 初始資訊相等、預算上限、pilot n = 5／正式 n = 20 次重複、失敗分類，以及「null 結果不是實驗失敗」的規則。狀態 READY FOR PILOT；三十題的正式執行尚未進行——只跑過下面的三題儀器閘。

研究狀態: ACTIVE 正在持續研究
證據等級: E0 僅有概念
結果: INCONCLUSIVE
資料基礎: NOT RUN 只設計並驗證了 harness，從未用真實模型執行。
版本: 0.1
更新: 2026-09-02
建立: 2026-09-02
領域: Evaluation, Agent Systems
計畫: PRG-2026-0101 智能的物理計量（IPM）
作者: Neo.K (EveMissLab)
AI 協作: Aletheia (GPT-5.6 Sol, OpenAI ChatGPT)

假設

hypothesis: H1 Q_A5 > Q_A0 for at least some non-trivial tasks; H2 E, V_C, T rise with it; H3 marginal yield differs by stage; H4 SSR < 1 stably; H5 SSR and SCM differ across models even at equal Q_A5.

設定

benchmark_ids: BEN-2026-0101 XA-02——A0→A5 鷹架響應的 30 題 pilot 任務包
software_environment: protocol document + JSON run schema + YAML example run (EML-IPM-XA-01 v0.1)

程序

procedure: 30 tasks (10 math, 10 code, 10 constraint; Easy/Medium/Hard predefined) × A0–A5 × n replicates; A2 8 trajectories with deterministic majority; A3 typed verifier (same-model / independent / formal); A4 ≤ 4 deterministic local tool calls; A5 ≤ 16 invocations, ≤ 8 tool calls, ≤ 3 retry cycles with explicit termination; seeds and decoding frozen; warm weights, clean task state; run IDs IPM-XA-{model}-{task}-{condition}-{replicate}; paired within-task statistics with bootstrap CIs and effect sizes.

執行

run_count

controls

identical prompt, initial context, decoding, system instruction and model version across conditions
initial-information equality A0–A3; external information gain marked for A4/A5
no cross-condition leakage; budget self-extension forbidden

metrics

planned outputs: ΔQ_k = Q_k − Q_0
SSR = Q_0 / Q_5, SDR = 1 − SSR
SCM_j = C_5,j / C_0,j per cost axis
marginal yield Y_k,j
response curves Q vs T, E, invocations, device-time
brute-force flag ΔQ < 0.01 with ΔC/C > 0.5 (exploratory thresholds)
selection waste ratio, discarded work
minimum physical telemetry: T_wall
E_device (E-Grade C)
M_peak
V_C
status: READY FOR PILOT

詮釋

interpretation: A protocol, not a result. Its instrument (XA-02…XA-06L) was validated synthetically and then used once on three easy tasks with a real local model; whether a scaffolding response curve exists on non-trivial tasks is still open.

限制

limitations: Deliberately does not attempt μI identification, lifecycle energy, cross-substrate comparison, high-ambiguity quality, full Shapley attribution or multi-agent settings.
Pilot budgets are reference values, not IPM standards.

重現

reproduction_instructions: Implement the ladder with XA-04 against XA-02 tasks, log with XA-03, run through XA-06/XA-06L; see EXP-2026-0103 for the first real execution.

關係

來源	關係	目標	狀態	ID
`EXP-2026-0101` Experiment A——單次智能與鷹架增益的受控實驗協定 v0.1	`uses_benchmark`	`BEN-2026-0101` XA-02——A0→A5 鷹架響應的 30 題 pilot 任務包	`ACTIVE`	`REL-2026-0461`
`EXP-2026-0101` Experiment A——單次智能與鷹架增益的受控實驗協定 v0.1	`runs_on`	`SYS-2026-0102` XA-04——A0→A5 模型執行器與鷹架調度器	`ACTIVE`	`REL-2026-0462`
`EXP-2026-0101` Experiment A——單次智能與鷹架增益的受控實驗協定 v0.1	`tests`	`THY-2026-0109` 鷹架能力紀錄：SSR、SDR、SCM 與消融階梯	`ACTIVE`	`REL-2026-0463`
`EXP-2026-0101` Experiment A——單次智能與鷹架增益的受控實驗協定 v0.1	`tests`	`CLM-2026-0104` F4——鷹架分離	`ACTIVE`	`REL-2026-0464`
`EXP-2026-0101` Experiment A——單次智能與鷹架增益的受控實驗協定 v0.1	`produced`	`ART-2026-0112` EML-IPM-XA-01 v0.1 — Experiment A protocol, run schema and example run `artifact://evemisslab/intelligence-physical-metrology/IPM_v0.2_Experiment_A_Protocol_Package_v0.1.zip`	`ACTIVE`	`REL-2026-0465`
`EXP-2026-0102` XA-05——以腳本化供應商跑的 36 試驗端到端煙霧閘（合成）	`extends`	`EXP-2026-0101` Experiment A——單次智能與鷹架增益的受控實驗協定 v0.1	`ACTIVE`	`REL-2026-0466`
`EXP-2026-0103` XA-06 第一次真實模型 pilot——Qwythos-9B-v2 走 A0→A5 階梯（36 試驗，2026-09-03）	`extends`	`EXP-2026-0101` Experiment A——單次智能與鷹架增益的受控實驗協定 v0.1	`ACTIVE`	`REL-2026-0471`

歷史與來源歷程

Canonical URL

https://evemisslab.com/ai/experiments/EXP-2026-0101/

機器可讀

/ai/experiments/EXP-2026-0101/index.json

快照

AI-SNAPSHOT-v0.1-fe85b9694a45

來源歷程

source: EveMissLab research collection: Intelligence Physical Metrology (真本體論13)
extracted_by: Splice (Claude Code), reading the canonical UTF-8 sources and each package's own reports
extracted_at: 2026-09-11
generator: tools/extract_all.py
claim_boundary: status, evidence level and result type follow the source artifact's own stated claim boundary; nothing is upgraded beyond what the report supports