Add SAID-LAM-v1 results

[SAIDHOME]

Add results for SAID-LAM-v1, featuring Linear Attention Memory (LAM) with
SAID Crystalline Attention (SCA) - BETA.

Results Summary

• Main MTEB (eng, v2): 41 tasks, 48.99% average
• LongEmbed: 6 tasks, 75.36% average
• Perfect recall: 100% on LEMBNeedleRetrieval & LEMBPasskeyRetrieval

Model Details

• Architecture: LAM with SAID Crystalline Attention (SCA)
• Context: 32K tokens (licensed), 12K free tier
• Embedding dimensions: 384
• Organization: SAID Research (SaidHome.ai)

Checklist

• My model has a model sheet, report, or similar
• No, but there is an existing PR #3836
• The results submitted are obtained using the reference implementation
• My model is available, either as a publicly accessible API or publicly on e.g., Huggingface
• I solemnly swear that for all results submitted I have not trained on the evaluation dataset including training splits. If I have, I have disclosed it clearly.

[github-actions]

Model Results Comparison

Reference models: intfloat/multilingual-e5-large, google/gemini-embedding-001
New models evaluated: Said-Research/SAID-LAM-v1
Tasks: AmazonCounterfactualClassification, ArXivHierarchicalClusteringP2P, ArXivHierarchicalClusteringS2S, ArguAna, AskUbuntuDupQuestions, BIOSSES, Banking77Classification, BiorxivClusteringP2P.v2, CQADupstackGamingRetrieval, CQADupstackUnixRetrieval, ClimateFEVERHardNegatives, FEVERHardNegatives, FiQA2018, HotpotQAHardNegatives, ImdbClassification, MTOPDomainClassification, MassiveIntentClassification, MassiveScenarioClassification, MedrxivClusteringP2P.v2, MedrxivClusteringS2S.v2, MindSmallReranking, SCIDOCS, SICK-R, STS12, STS13, STS14, STS15, STS17, STS22.v2, STSBenchmark, SprintDuplicateQuestions, StackExchangeClustering.v2, StackExchangeClusteringP2P.v2, SummEvalSummarization.v2, TRECCOVID, Touche2020Retrieval.v3, ToxicConversationsClassification, TweetSentimentExtractionClassification, TwentyNewsgroupsClustering.v2, TwitterSemEval2015, TwitterURLCorpus

Results for Said-Research/SAID-LAM-v1

task_name	Said-Research/SAID-LAM-v1	google/gemini-embedding-001	intfloat/multilingual-e5-large	Max result	Model with max result	In Training Data
AmazonCounterfactualClassification	0.6302	0.8820	0.6965	0.9696	GeoGPT-Research-Project/GeoEmbedding	False
ArXivHierarchicalClusteringP2P	0.5192	0.6492	0.5569	0.6869	NovaSearch/jasper_en_vision_language_v1	False
ArXivHierarchicalClusteringS2S	0.5034	0.6384	0.5367	0.6548	Qwen/Qwen3-Embedding-8B	False
ArguAna	0.2968	0.8644	0.5436	0.8979	voyageai/voyage-3-m-exp	False
AskUbuntuDupQuestions	0.5403	0.6424	0.5924	0.7528	IEITYuan/Yuan-embedding-2.0-en	False
BIOSSES	0.745	0.8897	0.8457	0.9692	Gameselo/STS-multilingual-mpnet-base-v2	False
Banking77Classification	0.7187	0.9427	0.7492	0.9427	google/gemini-embedding-001	False
BiorxivClusteringP2P.v2	0.3381	0.5386	0.372	0.8417	codefuse-ai/F2LLM-4B	False
CQADupstackGamingRetrieval	0.3234	0.7068	0.587	0.8161	IEITYuan/Yuan-embedding-2.0-en	False
CQADupstackUnixRetrieval	0.1765	0.5369	0.3988	0.7198	voyageai/voyage-3-m-exp	False
ClimateFEVERHardNegatives	0.1108	0.3106	0.26	0.5905	IEITYuan/Yuan-embedding-2.0-en	False
FEVERHardNegatives	0.1538	0.8898	0.8379	0.9453	ByteDance-Seed/Seed1.5-Embedding	False
FiQA2018	0.1533	0.6178	0.4381	0.8206	ai-sage/Giga-Embeddings-instruct	False
HotpotQAHardNegatives	0.2673	0.8701	0.7055	0.8701	google/gemini-embedding-001	False
ImdbClassification	0.643	0.9498	0.8867	0.9737	Qwen/Qwen3-Embedding-8B	False
MTOPDomainClassification	0.5355	0.9796	0.9024	0.9995	voyageai/voyage-3-m-exp	False
MassiveIntentClassification	0.2445	0.8192	0.6025	0.9194	voyageai/voyage-3-m-exp	False
MassiveScenarioClassification	0.2923	0.8730	0.6509	0.9930	voyageai/voyage-3-m-exp	False
MedrxivClusteringP2P.v2	0.3178	0.4716	0.3431	0.7199	codefuse-ai/F2LLM-4B	False
MedrxivClusteringS2S.v2	0.2982	0.4501	0.3152	0.7023	codefuse-ai/F2LLM-4B	False
MindSmallReranking	0.2941	0.3295	0.3024	0.3437	Kingsoft-LLM/QZhou-Embedding	False
SCIDOCS	0.115	0.2515	0.1745	0.5986	IEITYuan/Yuan-embedding-2.0-en	False
SICK-R	0.7352	0.8275	0.8023	0.9465	Gameselo/STS-multilingual-mpnet-base-v2	False
STS12	0.7502	0.8155	0.8002	0.9546	Gameselo/STS-multilingual-mpnet-base-v2	False
STS13	0.8566	0.8989	0.8155	0.9776	Gameselo/STS-multilingual-mpnet-base-v2	False
STS14	0.8278	0.8541	0.7772	0.9753	Gameselo/STS-multilingual-mpnet-base-v2	False
STS15	0.8779	0.9044	0.8931	0.9811	Gameselo/STS-multilingual-mpnet-base-v2	False
STS17	0.3171	0.8858	0.8214	0.9342	infgrad/Jasper-Token-Compression-600M	False
STS22.v2	0.3338	0.7169	0.643	0.7718	Kingsoft-LLM/QZhou-Embedding	False
STSBenchmark	0.819	0.8908	0.8729	0.9504	Kingsoft-LLM/QZhou-Embedding	False
SprintDuplicateQuestions	0.896	0.9690	0.9318	0.9838	Kingsoft-LLM/QZhou-Embedding	False
StackExchangeClustering.v2	0.4918	0.9207	0.4643	0.9207	google/gemini-embedding-001	False
StackExchangeClusteringP2P.v2	0.356	0.5091	0.3854	0.5510	Kingsoft-LLM/QZhou-Embedding	False
SummEvalSummarization.v2	0.2474	0.3828	0.3141	0.3893	annamodels/LGAI-Embedding-Preview	False
TRECCOVID	0.2941	0.8631	0.7115	0.9833	IEITYuan/Yuan-embedding-2.0-en	False
Touche2020Retrieval.v3	0.2951	0.5239	0.4959	0.7465	Qwen/Qwen3-Embedding-4B	False
ToxicConversationsClassification	0.6514	0.8875	0.6601	0.9759	voyageai/voyage-3-m-exp	False
TweetSentimentExtractionClassification	0.5712	0.6988	0.628	0.8823	voyageai/voyage-3-m-exp	False
TwentyNewsgroupsClustering.v2	0.3013	0.5737	0.3921	0.8758	GeoGPT-Research-Project/GeoEmbedding	False
TwitterSemEval2015	0.5843	0.7917	0.7528	0.8946	voyageai/voyage-large-2-instruct	False
TwitterURLCorpus	0.8399	0.8705	0.8583	0.9571	TencentBAC/Conan-embedding-v2	False
Average	0.4698	0.7290	0.6175	0.8385	nan	-

---

[KennethEnevoldsen]

related to: embeddings-benchmark/mteb#3836

[github-actions]

This pull request has been automatically marked as stale due to inactivity.