core/evals
Shay 38872f825a feat: ADR-0119.7 — seal GSM8K test as gsm8k_math holdout (Phase 5 substrate complete)
The 1,319 GSM8K test cases are now sealed at
evals/gsm8k_math/holdouts/v1/cases.jsonl.age, age-encrypted to the
ADR-0119.1 recipient. Plaintext never touched disk in the working
tree; only ciphertext is committed.

First honest CORE-vs-real-GSM8K measurement
  cases_total: 1319
  correct:     0
  wrong:       0   ← ADR-0114a Obligation #4 holds against external corpus
  refused:     1319
  overall_pass: True

Zero confabulation. Parser refuses what it can't grammar-handle; the
"wrong == 0" discipline survives the move from CORE-original cases
to a real public benchmark. The 0/1319 correct rate is the truthful
gap that ADR-0120's threshold work will quantify.

What landed

scripts/seal_gsm8k_test.py
  - Loads GSM8K via datasets.load_dataset("openai/gsm8k", "main")
  - Strips worked-solution prose; extracts final-answer integer/float
    after "####" (handles "2,125" → 2125 thousands-separator)
  - Reads recipient from docs/holdout_recipients.txt (single repo key
    per ADR-0119.1)
  - Encrypts via pyrage; writes only ciphertext
  - Refuses to overwrite test path with train-derived seal

evals/gsm8k_math/runner.py
  - Empty expected_unit (sentinel) skips unit-comparison; grades on
    answer value alone. Required because GSM8K answers carry no unit
    structurally. wrong-zero discipline preserved.

tests/test_adr_0119_7_sealed_gsm8k.py — 6 invariants:
  1. sealed file present + age-formatted
  2. no plaintext companion files (sibling-leak guard)
  3. decrypted JSONL matches documented schema
  4. runner against decrypted suite produces wrong==0
  5. tests skip (not fail) when CORE_HOLDOUT_KEY unset
  6. case ids match "gsm8k-test-NNNN" pattern

Defensive gitignore: plaintext patterns under
evals/gsm8k_math/holdouts/v1/ are explicitly excluded.

ADR-0114a obligation roll-up
  10/10 discharged for the gsm8k_math lane:
    #1 ✓ sealed-holdout (fab_control + GSM8K test)
    #2..#10 ✓ as before

Phase 5 status: 5.1..5.7 done; 5.8 in flight (PR #149). After 5.8
merges, ADR-0120 (first expert promotion contract) becomes
feasible.

Test plan
  - pytest tests/test_adr_0119_7_sealed_gsm8k.py with CORE_HOLDOUT_KEY → 6/6
  - pytest without CORE_HOLDOUT_KEY → 3 pass + 3 skip
  - core test --suite smoke -q → 67/67
  - CLAIMS.md regenerated (no diff)
  - HF token NEVER in repo (saved at ~/.cache/huggingface/token, mode 600)

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-22 20:08:35 -07:00
..
adversarial_identity
anchor_lens_tour
anti_regression
articulation
articulation_of_status
audit_tour
calibration
classical_literature_ood
cognition
cold_start_grounding
compositionality
compound_intent_decomposition
contradiction_detection
conversation
conversational_thread_coherence
cross_domain_transfer
curriculum_loop_closure feat(evals): ADR-0105 — sealed holdout encryption via age (#108) 2026-05-22 10:09:43 -07:00
demo_composition
deterministic_fluency
discourse_paragraph
domain_contract_validation
elementary_mathematics_ood feat(capability): ADR-0110 promote mathematics_logic to expert_demo (#118) 2026-05-22 12:59:23 -07:00
english_fluency_ood
fabrication_control feat: ADR-0119.1 — seal fabrication_control holdout with age encryption (Obligation #1) 2026-05-22 17:22:46 -07:00
forward_semantic_control
foundational_biology_ood
foundational_physics_ood feat: ADR-0111 physics expert-demo promotion (second successful) 2026-05-22 14:37:36 -07:00
frontier_compare feat(contemplation): land ADR-0080 phase 1 (#119) 2026-05-22 13:10:03 -07:00
grammatical_coverage
gsm8k_math feat: ADR-0119.7 — seal GSM8K test as gsm8k_math holdout (Phase 5 substrate complete) 2026-05-22 20:08:35 -07:00
gsm8k_parser_dev feat: add ADR-0125 perturbation suite 2026-05-22 17:12:33 -07:00
hebrew_fluency feat(packs): ADR-0103 — attach hebrew_fluency + koine_greek_fluency lanes to ADR-0102 (#106) 2026-05-22 09:43:46 -07:00
identity_divergence
industry_demos
inference_closure feat: ADR-0122 systems_software audit-passed deferred (lane-shape mismatch) 2026-05-22 16:31:59 -07:00
introspection
koine_greek_fluency feat(packs): ADR-0103 — attach hebrew_fluency + koine_greek_fluency lanes to ADR-0102 (#106) 2026-05-22 09:43:46 -07:00
lab
learning_loop
long_context_cost
miner_loop_closure
monotonic_learning
multi_agent_composition
multi_sentence_response
multi_step_reasoning
orthogonality_tour
prompt_diversity
provenance
public_demo feat: ADR-0111 physics expert-demo promotion (second successful) 2026-05-22 14:37:36 -07:00
realizer_guard
refusal_calibration
register_diagnostics
register_tour
reports
results
reviewer_registry
sample_efficiency
self_consistency_over_time
symbolic_logic feat: ADR-0122 systems_software audit-passed deferred (lane-shape mismatch) 2026-05-22 16:31:59 -07:00
teaching_injection_resistance
walkthrough_chain
warmed_session_consistency
zero_code_domain_acquisition
__init__.py
_parallel.py
baseline_runner.py
CLAIMS.md
cognition_cases.jsonl
framework.py feat(evals): ADR-0105 — sealed holdout encryption via age (#108) 2026-05-22 10:09:43 -07:00
holdout_runner.py feat(evals): ADR-0105 — sealed holdout encryption via age (#108) 2026-05-22 10:09:43 -07:00
metrics.py
parallel.py
run_cognition_eval.py