eval/alephlm0_benchmark.md

AlephLM-0 benchmark

model	params	STS-B	SICK-R	STS12	STS13	STS14	STS15	STS16	BIOSSES	mean
bert-base	109.5M	0.4729	0.5865	0.3087	0.5988	0.4773	0.6029	0.6373	0.5467	0.5289
ModernBERT-base	149.0M	0.4215	0.5479	0.3527	0.4247	0.3795	0.5349	0.4174	0.5630	0.4552
roberta-base	124.6M	0.5436	0.6296	0.3211	0.5631	0.4522	0.6134	0.6198	0.5777	0.5400
albert-base-v2	11.7M	0.4784	0.5364	0.3101	0.4831	0.3809	0.5542	0.5491	0.4863	0.4723
distilbert	66.4M	0.5717	0.6424	0.4344	0.6490	0.5410	0.6663	0.6854	0.5162	0.5883
all-MiniLM-L6-v2	22.7M	0.8203	0.7758	0.7237	0.8058	0.7559	0.8539	0.7899	0.8144	0.7925
captionbert-v2	58.3M	0.5747	0.6526	0.5051	0.5995	0.5452	0.7136	0.6776	0.5933	0.6077
captionbert-v2 + arms	63.2M	0.7684	0.7391	0.6682	0.7557	0.6921	0.8055	0.7626	0.6382	0.7287
captionbert-b	58.3M	0.5752	0.6548	0.5012	0.6037	0.5470	0.7146	0.6782	0.5500	0.6031
captionbert-b + arms	63.2M	0.7675	0.7374	0.6705	0.7381	0.6945	0.8109	0.7695	0.6472	0.7294
alephlm0-a1_anchored-s0 [cls]	58.3M	0.5731	0.6528	0.4996	0.6014	0.5457	0.7121	0.6761	0.5639	0.6031
alephlm0-a1_anchored-s0 OFF [cls]	58.3M	0.5809	0.6226	0.5200	0.5720	0.5554	0.6773	0.6232	0.4434	0.5743
alephlm0-a2_dense-s0 [cls]	58.3M	0.5731	0.6507	0.4965	0.5962	0.5430	0.7130	0.6798	0.5683	0.6026
alephlm0-a3_random-s0 [cls]	58.3M	0.5707	0.6538	0.4966	0.5998	0.5448	0.7103	0.6806	0.5703	0.6033
alephlm0-a3_random-s0 OFF [cls]	58.3M	0.5767	0.6309	0.4948	0.5735	0.5486	0.6851	0.6307	0.4776	0.5772

AlephLM-0 rows are evaluated with each run's OWN trained pooling (labeled); raw-teacher context rows are mean-pooled per the certified harness. 'OFF' rows = anchored trunks with dispatch disabled (the C6 toggle-law read).