Commit graph

29 commits

Author SHA1 Message Date
twentyOne2x
2712e7fbb5 Harden Leo benchmark semantic invariants 2026-07-16 11:38:51 +02:00
twentyOne2x
3b0e1bc296 Harden Leo DB contract correction path 2026-07-16 10:38:58 +02:00
twentyOne2x
31782d9b38 Accept semantic Leo benchmark paraphrases 2026-07-16 09:20:52 +02:00
twentyOne2x
b59bee8844 Fix Leo restart truth and read command reliability 2026-07-16 07:14:31 +02:00
twentyOne2x
021aaf6f09 Harden blind scoring for natural Leo replies 2026-07-16 06:13:53 +02:00
twentyOne2x
8cbda55435 Preserve subject and memory boundaries in Leo replies 2026-07-16 05:59:09 +02:00
twentyOne2x
3cbbbfaa18 Harden broad Leo database reasoning contracts 2026-07-16 05:44:21 +02:00
twentyOne2x
813d4cd91d Harden blinded Leo reasoning repair gate 2026-07-16 05:19:50 +02:00
twentyOne2x
4daa7853fe Integrate evidence-backed Leo reasoning candidate 2026-07-16 01:14:00 +02:00
twentyOne2x
27fbfa49a9
Accept receipt-bound natural answers in the Leo OOS scorer (#172)
* Make OOS scoring accept receipt-bound natural answers

* Test short-ID causal attribution in OOS ablation

* Close OOS provenance and schema qualifier gaps
2026-07-15 23:44:46 +02:00
twentyOne2x
8e198a791b Ground Leo reasoning in bounded live retrieval 2026-07-15 21:21:06 +02:00
twentyOne2x
8a4325f42a Harden OOS benchmark provenance and leakage checks 2026-07-15 12:06:35 +02:00
twentyOne2x
da65e05316 Ground Leo reasoning in bounded live retrieval 2026-07-15 05:30:10 +02:00
twentyOne2x
ee48f82f06
Stage source artifacts as reviewable KB proposals (#123)
Some checks are pending
CI / lint-and-test (push) Waiting to run
2026-07-13 20:38:18 +02:00
twentyOne2x
73f466f899
Make broad Leo replies deterministic (#121)
Some checks are pending
CI / lint-and-test (push) Waiting to run
2026-07-13 20:07:51 +02:00
twentyOne2x
5d135d8818
Make broad Leo answers database-composed (#120)
Some checks are pending
CI / lint-and-test (push) Waiting to run
2026-07-13 19:53:25 +02:00
twentyOne2x
61c1fa9165
Accept explicit source-intake prohibitions (#117)
Some checks are pending
CI / lint-and-test (push) Waiting to run
2026-07-13 13:35:34 +02:00
twentyOne2x
2264ad1f2b
Enforce DB-composed Leo responses before delivery (#116)
Some checks are pending
CI / lint-and-test (push) Waiting to run
2026-07-13 13:27:54 +02:00
twentyOne2x
144e619eee Keep Leo composition replies within budget 2026-07-13 12:58:24 +02:00
twentyOne2x
033e251fd0 Disambiguate governance rules in Leo composition 2026-07-13 12:52:50 +02:00
twentyOne2x
e3af44772e Ground Leo composition in live Postgres constraints 2026-07-13 12:46:50 +02:00
twentyOne2x
d9d799d725 Drive Leo answers from live database contracts 2026-07-13 12:20:30 +02:00
twentyOne2x
17c7875129 Make Leo honor intake and reply boundaries 2026-07-13 12:06:04 +02:00
twentyOne2x
c7bbe37116 Correct Leo source intake and mixed packet boundaries 2026-07-13 11:55:35 +02:00
twentyOne2x
ac41fbb958 Expose proposal readiness and current KB truth 2026-07-13 11:43:08 +02:00
twentyOne2x
8c20613fd0 Expand Leo database grounding and blind behavior suite 2026-07-13 10:50:42 +02:00
twentyOne2x
9f4e4cac0e Pin Leo composition to the shipped apply contract 2026-07-13 09:58:32 +02:00
twentyOne2x
64c35a7589 Harden Leo against schema and receipt drift 2026-07-13 09:17:28 +02:00
twentyOne2x
c6c8a5eb31 Enforce m3taversal identity and schema truth 2026-07-13 08:55:46 +02:00
Renamed from tests/test_working_leo_cory_oos_benchmark.py (Browse further)