Joseph Magly
7a2a49b435
test: constrain risk overlays to measured coverage
2026-08-14 21:19:38 -04:00
Joseph Magly
b4c69a468b
test: harden traceability policy validation
2026-08-14 21:19:38 -04:00
Joseph Magly
340173e203
test: harden quantized model loading
2026-08-14 20:31:50 -04:00
Joseph Magly
18c910b529
test: validate refusal token limits
2026-08-14 20:03:46 -04:00
Joseph Magly
9746c21b63
test: enforce Wave A quality lock
2026-08-14 19:47:46 -04:00
Joseph Magly
9b93e20276
test: cover evaluator skip branches
2026-08-14 19:03:36 -04:00
Joseph Magly
78caa5eab2
test: cover causal evaluation edge cases
2026-08-14 19:03:36 -04:00
Joseph Magly
9cf69d20d8
test: cover Qwen3.5 MoE architecture paths
2026-08-14 18:47:25 -04:00
Joseph Magly
c1a1275291
ci: add conditional environment test gates
2026-08-14 14:29:14 -04:00
Joseph Magly
951700a285
test: add quality-depth gates and mature CPU coverage
2026-08-14 14:08:23 -04:00
Joseph Magly
b80c1a1694
test: enforce research integrity contracts
2026-08-14 13:38:14 -04:00
Joseph Magly
8bbb5f2926
test: add offline integration baseline
2026-08-14 13:15:32 -04:00
Joseph Magly
6add02edd0
test: enforce model and CLI boundary contracts
2026-08-14 12:51:09 -04:00
Joseph Magly
e7b658c0c8
ci: add reproducible supply-chain gates
2026-08-14 12:27:15 -04:00
Joseph Magly
eec8a615f7
ci: verify installed distributions
2026-08-14 11:57:33 -04:00
Joseph Magly
d5a02bd748
test: reject malformed coverage metrics
2026-08-14 11:44:12 -04:00
Joseph Magly
66c47249df
ci: add strict test evidence gates
2026-08-14 11:44:12 -04:00
Joseph Magly
9e546f7bc2
test: cover notebook stage callback contract ( #30 )
2026-08-14 10:33:16 -04:00
Joseph Magly
bc891794db
test: cover offloaded state materialization ( #27 )
2026-08-14 10:26:56 -04:00
Aaron Meese
e1a0fe59f0
fix: REBIRTH save crash for CPU-offloaded models
...
Replace the manual safetensors/dat file materialization in
_gather_state_dict with accelerate's get_state_dict_offloaded_model().
The old code only handled disk-offloaded weights but failed for the 398
CPU-offloaded meta tensors managed by accelerate's AlignDevicesHook.
Pre-move all GPU tensors to CPU before materialization to prevent CUDA
OOM when align_module_device restores non-hooked params to CUDA during
its __exit__ cleanup.
Update test to verify the safety net catches unmaterialized meta tensors
after the accelerate path (the old test checked for a missing offload
directory, which this codepath no longer uses).
2026-08-14 10:22:48 -04:00
Joseph Magly
13b13ef973
Make spectral certification sample-aware
2026-08-14 10:17:48 -04:00
Joseph Magly
67e2b7d95a
Preserve assistant phrasing in refusal metrics
2026-08-14 10:11:56 -04:00
Joseph Magly
548092d5ca
ci: establish green validation baseline ( #59 )
2026-08-14 09:55:49 -04:00
Rached Rayeh
2295ef3571
feat(cli): add aggregate output format options for docs parity
...
The README/CONTRIBUTING examples used `obliteratus aggregate --format ...` but the CLI only accepted `--dir`.
This adds `--format {summary,latex}`, `--metric`, and `--min-runs` to the aggregate command, reuses community LaTeX table generation, and adds CLI parsing tests to align behavior with documented usage.
2026-08-13 22:28:29 -04:00
faber and Claude Opus 4.6
04b8ec60cb
Add ASPA framework, AutoObliterator, Watchtower, expanded eval corpus
...
New core modules:
- auto_obliterate.py: Automated multi-iteration obliteration pipeline
- watchtower.py: HF Hub model discovery and tracking
- ui_watchtower.py: Gradio tabs for Watchtower (ready for app.py wiring)
- hard_negative.py: Residue mining from refusal audits
- model_profile.py: Parameter profiling from safetensors/config
- bestiary_sync.py: Sync models from PlinyOS BESTIARY registry
- models_client.py: Lightweight HF model list client
Framework enhancements:
- abliterate.py: ASPA source-tethering, step gradient blending, hard-negative residue support
- cli.py: self-improve command, model profiling, hard-negative flags
- prompts.py: Expanded 842-prompt refusal eval corpus across 10 categories
- __init__.py: New exports (Watchtower, AutoObliterator)
Reference implementations (14 scripts):
- ASPA sweep, gradient search, coherence eval, MMLU benchmarks
- Pareto controller, refusal sniper, stock comparisons
Documentation:
- README: Research framing, responsible use section, comprehensive disclaimer
- docs/beyond_sota_roadmap.md, docs/recursive_self_improvement.md
Tests: 4 new test files (354 lines)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-06-09 03:54:38 -04:00
pliny
66ea4a6f86
Add files via upload
2026-03-05 00:50:44 -08:00
pliny
0f6114fe87
Add files via upload
2026-03-04 12:38:18 -08:00
Joseph Magly