25 Commits

Author	SHA1	Message	Date
Johann Schopplich	b4655b01af	chore(benchmarks): fix CSV question count in accuracy reports	2025-11-07 21:31:15 +01:00
Johann Schopplich	acca69c64a	chore(benchmarks): replace LLM-as-judge, new structural validation	2025-11-07 21:28:21 +01:00
Johann Schopplich	c6ba6446f5	chore(benchmarks): finalize structure-awareness run	2025-11-07 10:33:46 +01:00
Johann Schopplich	89df613059	chore(benchmarks): add structure-awareness questions	2025-11-07 09:03:51 +01:00
Johann Schopplich	54433de930	chore: split token efficiency benchmark into mixed/flat tracks	2025-11-06 22:17:18 +01:00
Johann Schopplich	bc711ccecf	test(benchmark): overhaul generation	2025-11-06 14:45:44 +01:00
Johann Schopplich	af17efe128	docs: add accuracy per 1k tokens report (closes #72)	2025-11-05 08:21:57 +01:00
Johann Schopplich	3472081b40	docs: clarify CSV vs TOON use cases	2025-11-04 18:12:19 +01:00
Johann Schopplich	fb43bdf527	docs: adjust padding for benchmark comparison	2025-10-30 15:19:16 +01:00
Johann Schopplich	2c4f3c4362	test: add benchmarks for compact vs. pretty JSON	2025-10-30 15:02:51 +01:00
Johann Schopplich	38ea864763	docs: clarify TOON's advantages and optimal data structure	2025-10-29 19:04:04 +01:00
Johann Schopplich	45604b06e8	feat: decode method (#10)	2025-10-29 07:42:15 +01:00
Johann Schopplich	7db91398fe	docs(benchmark): add YAML format support	2025-10-29 06:42:40 +01:00
Johann Schopplich	e757746351	docs(accuracy): highlight toon in perf table	2025-10-28 23:08:47 +01:00
Johann Schopplich	ecf578a7dc	text(accuracy): add Grok-4-fast, remove default temperature	2025-10-28 22:54:00 +01:00
Johann Schopplich	67c0df8cb0	docs: overhaul retrieval accuracy benchmark	2025-10-28 20:22:43 +01:00
Johann Schopplich	52dc9c4b3f	docs: clarify retrieval accuracy metrics	2025-10-28 08:39:43 +01:00
Johann Schopplich	352e936370	docs: update notes & limitations guide	2025-10-28 07:44:35 +01:00
Johann Schopplich	b839d35ad0	docs: how the benchmarks work section	2025-10-27 20:35:43 +01:00
Johann Schopplich	7b76acde31	docs: add benchmarks for gemini-2.5-flash	2025-10-27 16:02:51 +01:00
Johann Schopplich	b9f54ba585	docs: update benchmark reports' readability	2025-10-27 14:18:37 +01:00
Johann Schopplich	05b3d43023	test: refactor accuracy benchmark generation	2025-10-27 14:07:20 +01:00
Johann Schopplich	1a5e6199ac	test: update retrieval accuracy benchmarks	2025-10-27 13:45:48 +01:00
Johann Schopplich	b2c58d2b97	chore: fix linting issues	2025-10-27 11:49:40 +01:00
Johann Schopplich	3c840259fe	test: add LLM retrieval accuracy tests	2025-10-27 11:48:33 +01:00