EVEMISSLAB
English

實驗EXP-2026-0009v0.1

PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何

加入非線性約束場家族 N3 約束競爭,以及專門用來破壞加性/log-odds 映射的對抗性幾何,所有 v0.1 門檻不變。N3 在三種幾何下都達 Level 3;N2 在對抗幾何下一致度 0.8546、更新/干預誤差仍小,但 held-out 表徵 JS 0.053327 越過 0.05 門檻,歸為僅行為收斂。線性診斷下的獨立家族:{N0, N1}、{N2}、{N3};三種幾何下都穩健的:兩個。

研究狀態
STABLE 目前的研究結論相對穩定
證據等級
E3 重複實驗
結果
MIXED
資料基礎
SYNTHETIC 合成數據與理論推理。現在很多人把合成數據當成真的;這個實驗室刻意反過來說——在真正的混合模型出現之前,推論就只是推論,理論上可能不等於實際上可能。
版本
0.1
更新
2026-09-09
建立
2026-09-09
領域
Model Representation, Evaluation, Formal AI
計畫
PRG-2026-0001 自適應世界狀態系統的第一原理框架
作者
Neo.K (EveMissLab)
AI 協作
Sol (GPT-5.6, OpenAI ChatGPT)

假設

hypothesis
A structurally distinct third family also converges, and convergence survives an evidence geometry designed against additive mappings.

設定

model_ids
dataset_ids
benchmark_ids
software_environment
Python; deterministic seeded generators; no network, no LLM.

執行

run_count
3
random_seeds
  • 20260908
  • 8 secondary seeds (smaller scale)
  • 5 adversarial seeds near primary scale
controls
  • shuffled-target mapping
  • constant prediction
  • broken composition (v0.6+)
  • target-local refit (v0.7+)
metrics
verdict
THIRD INDEPENDENT FAMILY CONVERGES; ROBUST PACC-A GATE REMAINS OPEN
adversarial
N0
agreement
0.90818
heldout_js
0.023921
level
3
N2
agreement
0.85455
heldout_js
0.053327
level
1
N3
agreement
0.90162
heldout_js
0.014454
level
3
five_seed_adversarial_level3_rate
N0
1.0
N1
1.0
N2
0.6
N3
1.0
independent_convergent_families
2
pacc_a
false

詮釋

interpretation
Probability-like convergence has a nontrivial basin, not a demonstrated universal attractor: the N3 result weakens the 'disguised additive accumulator' explanation, and N2 shows convergence is not guaranteed for every architecture under every geometry.

限制

limitations
  • Independence is supported under the preregistered linear diagnostic only, not proof of deep nonlinear nonequivalence.
  • N2's boundary is adversarial sensitivity around the threshold, not a universal phase boundary.

重現

reproduction_instructions
Extract the version's FINAL bundle; python -m pytest -q; run the version's primary script with the recorded seed; docs/PACC_LAB_v0.N_RESULTS.md and docs/EXPERIMENT_PROTOCOL_v0.N.md are inside the bundle.

結果

來源關係目標狀態ID
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何producesRST-2026-0007 v0.2 對抗幾何:N3 收斂、N2 越過表徵門檻ACTIVEREL-2026-0180

記錄欄位

completed_at
2026-09-09

關係

來源關係目標狀態ID
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何runs_onSYS-2026-0002 PACC-Lab——收斂猜想的微型實驗室ACTIVEREL-2026-0169
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何uses_benchmarkBEN-2026-0001 PACC 微型實驗室協定 v0.1(凍結門檻)ACTIVEREL-2026-0170
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何uses_datasetDAT-2026-0001 PACC 合成證據世界ACTIVEREL-2026-0171
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何uses_modelMOD-2026-0001 N0——帶號支持ACTIVEREL-2026-0172
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何uses_modelMOD-2026-0002 N1——序數錦標賽ACTIVEREL-2026-0173
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何uses_modelMOD-2026-0003 N2——帶號圖ACTIVEREL-2026-0174
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何uses_modelMOD-2026-0004 N3——約束競爭ACTIVEREL-2026-0175
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何uses_modelMOD-2026-0005 Bayesian 參考(精確後驗/Beta-Bernoulli/聯合共同因/HMM)ACTIVEREL-2026-0176
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何testsTHY-2026-0005 PACC 猜想——四層收斂階梯ACTIVEREL-2026-0177
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何extendsEXP-2026-0008 PACC-Lab v0.1——E0–E4:第一個微型見證ACTIVEREL-2026-0178
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何producedART-2026-0022 PACC-Lab v0.2 N3 Adversarial FINAL artifact://evemisslab/adaptive-epistemic-systems/PACC-Lab_v0.2_N3_Adversarial_FINAL.zipACTIVEREL-2026-0179
EXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何producesRST-2026-0007 v0.2 對抗幾何:N3 收斂、N2 越過表徵門檻ACTIVEREL-2026-0180
EXP-2026-0010 PACC-Lab v0.3——無 oracle 的來源可靠度學習extendsEXP-2026-0009 PACC-Lab v0.2——第三個家族(N3)與對抗性證據幾何ACTIVEREL-2026-0191

歷史與來源歷程

Canonical URL
https://evemisslab.com/ai/experiments/EXP-2026-0009/
機器可讀
/ai/experiments/EXP-2026-0009/index.json
快照
AI-SNAPSHOT-v0.1-fe85b9694a45
來源歷程
source
EveMissLab research collection: Adaptive Epistemic Systems (真本體論13)
extracted_by
Splice (Claude Code), reading the canonical UTF-8 sources and each lab's own result reports
extracted_at
2026-09-11
generator
tools/extract_aes/extract.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