track	probe	representation	tag	modality	text_in_training	in_v0	n_genes	generated_text_share	index_0-10	index_0-10_lo	index_0-10_hi	rank_0-10	rank_0-10_lo	rank_0-10_hi	index_all	index_all_lo	index_all_hi	rank_all	index4_0-10	index4_all	T1_GO_leakfree_0-10	T1_all	T2_assay_family_grouped_0-10	T2_all	T3_time_travel_0-5	T3_all	T4_coessential_0-10	T4_all	S_provenance_sensitivity_1-10
O	linear	GB.Protein-16B + rich leak-free text (concatenation)	fusion_p16b_refseq	mixture		False	19256	0.0000	47.0887	46.3240	47.8080	1.0000	1.0000	2.0000	50.4260	49.9968	50.8745	1.0000			0.9212	0.9081	0.6578	0.7203			0.6273	0.6280	
O	linear	ProTrek-650M	protrek	text-aligned protein	UniProt function text (Swiss-Prot 2024)	False	19133		46.9633	46.2118	47.6505	2.0000	1.0000	2.0000	48.7690	48.3287	49.2030	2.0000			0.9037	0.8909	0.6979	0.7486			0.6029	0.5921	
O	linear	ProtST-ESM1b	protst	text-aligned protein	UniProt function text (Swiss-Prot 2022)	False	19133		46.3322	45.5652	47.0450	3.0000	2.0000	3.0000	48.0305	47.5909	48.4629	3.0000	51.5446	54.2862	0.8918	0.8778	0.7293	0.7730	0.8359	0.8653	0.5739	0.5697	
O	linear	GB.Protein-16B	protein16b	protein		True	19133		45.2621	44.5114	46.0354	4.0000	4.0000	4.0000	46.1877	45.7583	46.6342	7.0000	47.9643	51.2877	0.8753	0.8574	0.6763	0.7192	0.7804	0.8329	0.6274	0.6163	
O	linear	ESM2-650M	esm2_650m	protein		True	19133		44.4197	43.6560	45.1241	5.0000	5.0000	6.0000	44.9356	44.4783	45.3699	9.0000	47.1706	50.6567	0.8792	0.8601	0.6905	0.7288	0.7771	0.8391	0.5967	0.5851	
O	linear	ESM1b	esm1b	protein		False	19133		43.9140	43.1473	44.7076	6.0000	5.0000	6.0000	44.4558	44.0275	44.9037	10.0000	46.3930	49.6030	0.8733	0.8523	0.6736	0.7138	0.7691	0.8252	0.6118	0.6007	
O	linear	GenePT text-embedding-3-large, as released	genept_3large	text		True	18945	0.3088	43.0429	42.2760	43.7616	7.0000	7.0000	9.0000	46.0052	45.5492	46.4623	8.0000			0.9187	0.8979	0.6557	0.7126			0.5713	0.5796	
O	linear	Rich leak-free text, Qwen3-Emb-8B	text_refseq	text		True	19256	0.0000	42.8099	42.0122	43.5853	8.0000	7.0000	9.0000	47.8121	47.3927	48.2568	4.0000			0.9153	0.9079	0.6509	0.7221			0.5759	0.5872	
O	linear	UniProt function text only, Qwen3-Emb-8B	text_uniprot	text		False	19256	0.0000	42.4560	41.6649	43.1860	9.0000	8.0000	9.0000	46.5648	46.1274	47.0546	6.0000			0.9162	0.9070	0.6466	0.7064			0.5740	0.5851	
O	linear	GenePT text as found, Qwen3-Emb-8B	t23a__qwen8b	text		True	19256	0.3039	41.4975	40.7056	42.2108	10.0000	10.0000	11.0000	44.4337	43.9621	44.8731	11.0000			0.9044	0.8833	0.6533	0.7125			0.5647	0.5707	
O	linear	GenePT ada-002, NCBI text as released	genept_ada	text		True	18942	0.3089	41.4426	40.6632	42.1927	11.0000	10.0000	11.0000	42.5402	42.0802	43.0152	12.0000	41.0867	46.1056	0.8992	0.8718	0.6606	0.7010	0.7001	0.7840	0.5618	0.5653	
O	linear	GB.RNA-1.6B	rna1b6	rna		False	19228		39.9036	39.2001	40.6487	12.0000	12.0000	13.0000	40.4368	39.9857	40.8681	15.0000	42.5765	44.4188	0.8003	0.7852	0.6707	0.7085	0.7530	0.7818	0.6275	0.6129	
O	linear	Rich leak-free text, names masked, Qwen3-Emb-8B	text_refseq_nomenfree	text		True	19256	0.0000	39.2205	38.3554	40.1124	13.0000	12.0000	13.0000	46.8981	46.4669	47.3078	5.0000			0.8588	0.8875	0.6549	0.7292			0.5746	0.5867	
O	linear	Geneformer V2-316M	geneformer	single-cell		True	19229		36.5876	35.7530	37.3430	14.0000	14.0000	15.0000	37.7303	37.2947	38.1410	17.0000	35.7601	37.7898	0.7275	0.7309	0.7333	0.7455	0.6664	0.6898	0.5879	0.5896	
O	linear	scELMo, GPT-3.5-written descriptions	scelmo_gpt35	text		True	19231	0.0000	36.0799	35.2309	36.9541	15.0000	14.0000	16.0000	42.2789	41.7884	42.7230	13.0000	38.7203	47.5686	0.8393	0.8552	0.6428	0.7072	0.7332	0.8172	0.5591	0.5718	
O	linear	Gene symbol only, Qwen3-Emb-8B	text_symbol	text		True	19256	0.0000	35.6011	34.7085	36.4037	16.0000	15.0000	17.0000	39.3011	38.8193	39.7314	16.0000			0.8539	0.8480	0.6232	0.6758			0.5569	0.5657	
O	linear	GB.Cell-100M	cell_base	single-cell		True	18593		34.6975	33.9094	35.4708	17.0000	16.0000	18.0000	37.6178	37.1849	38.0457	18.0000	34.9775	39.1384	0.7079	0.7294	0.7327	0.7567	0.6791	0.7185	0.5798	0.5781	
O	linear	GenePT text, generated summaries removed, Qwen3-Emb-8B	t23l__qwen8b	text		True	19256	0.0000	34.1856	33.2807	35.0808	18.0000	17.0000	18.0000	40.7003	40.2273	41.1867	14.0000			0.8325	0.8490	0.6242	0.7085			0.5561	0.5531	
O	linear	Qwen2.5-7B-written descriptions, Qwen3-Emb-8B	llm_qwen25	text		True	19256	0.0000	31.1909	30.3485	32.0474	19.0000	19.0000	20.0000	35.7360	35.2494	36.2079	19.0000			0.8188	0.8202	0.5997	0.6585			0.5493	0.5574	
O	linear	GB.DNA-300M	dna300m	dna		False	19236		30.9740	30.1493	31.8018	20.0000	19.0000	20.0000	29.3018	28.8360	29.7731	20.0000	30.4949	30.0086	0.7045	0.6832	0.6779	0.6852	0.6453	0.6606	0.5823	0.5711	
F	linear	pubmedbert-base-embeddings	t23l__pubmedbert	text		True	19256	0.0000	34.6358						41.3726						0.8220	0.8461	0.6409	0.7235			0.5567	0.5510	0.0896
F	linear	Qwen3-Embedding-8B	t23l__qwen8b	text		True	19256	0.0000	34.1856						40.7003						0.8325	0.8490	0.6242	0.7085			0.5561	0.5531	0.0779
F	linear	all-mpnet-base-v2	t23l__mpnet	text		True	19256	0.0000	33.3400						39.0574						0.8196	0.8353	0.6258	0.6998			0.5548	0.5508	0.0884
F	linear	Qwen3-Embedding-0.6B	t23l__qwen06b	text		True	19256	0.0000	32.9254						38.7130						0.8218	0.8382	0.6206	0.6980			0.5515	0.5445	0.0875
F	linear	TF-IDF (1,024-d SVD)	t23l__tfidf	text		True	19256	0.0000	32.4682						34.9941						0.8300	0.8186	0.6125	0.6740			0.5445	0.5323	0.0429
F	linear	bge-large-en-v1.5	t23l__bge	text		True	19256	0.0000	32.3062						39.1068						0.8208	0.8390	0.6113	0.6970			0.5525	0.5506	0.0921
O	mlp	GB.Protein-16B + rich leak-free text (concatenation)	fusion_p16b_refseq	mixture		False	19256	0.0000	58.2472	57.5708	58.8592	1.0000	1.0000	1.0000	59.5076	59.1266	59.8912	1.0000			0.9696	0.9524	0.7769	0.8122			0.6273	0.6280	
O	mlp	GB.Protein-16B	protein16b	protein		True	19133		54.3180	53.6711	54.9358	2.0000	2.0000	3.0000	53.6411	53.2473	54.0557	5.0000	61.6105	61.6317	0.9446	0.9258	0.7428	0.7626	0.9174	0.9280	0.6274	0.6163	
O	mlp	ProTrek-650M	protrek	text-aligned protein	UniProt function text (Swiss-Prot 2024)	False	19133		53.9396	53.2892	54.5601	3.0000	2.0000	3.0000	54.0297	53.6028	54.4430	3.0000			0.9596	0.9430	0.7467	0.7754			0.6029	0.5921	
O	mlp	ESM1b	esm1b	protein		False	19133		53.2482	52.6135	53.9149	4.0000	4.0000	4.0000	52.2364	51.8291	52.6383	8.0000	60.7538	60.5556	0.9449	0.9276	0.7420	0.7552	0.9164	0.9276	0.6118	0.6007	
O	mlp	ESM2-650M	esm2_650m	protein		True	19133		52.1481	51.5106	52.7853	5.0000	5.0000	7.0000	51.3872	50.9882	51.7599	11.0000	59.8228	60.0459	0.9465	0.9290	0.7391	0.7567	0.9142	0.9301	0.5967	0.5851	
O	mlp	Rich leak-free text, Qwen3-Emb-8B	text_refseq	text		True	19256	0.0000	52.1422	51.4490	52.8430	6.0000	5.0000	7.0000	55.1175	54.7249	55.5092	2.0000			0.9674	0.9500	0.7388	0.7896			0.5759	0.5872	
O	mlp	UniProt function text only, Qwen3-Emb-8B	text_uniprot	text		False	19256	0.0000	51.6084	50.8875	52.2950	7.0000	6.0000	8.0000	53.9790	53.5525	54.3932	4.0000			0.9668	0.9488	0.7333	0.7758			0.5740	0.5851	
O	mlp	ProtST-ESM1b	protst	text-aligned protein	UniProt function text (Swiss-Prot 2022)	False	19133		51.5732	50.9354	52.2209	8.0000	6.0000	8.0000	51.7584	51.3270	52.1344	10.0000	59.6739	60.2700	0.9493	0.9365	0.7504	0.7702	0.9199	0.9290	0.5739	0.5697	
O	mlp	GenePT text-embedding-3-large, as released	genept_3large	text		True	18945	0.3088	50.8738	50.2307	51.5643	9.0000	9.0000	9.0000	52.5678	52.1353	52.9828	7.0000			0.9641	0.9423	0.7277	0.7666			0.5713	0.5796	
O	mlp	GenePT text as found, Qwen3-Emb-8B	t23a__qwen8b	text		True	19256	0.3039	50.0408	49.3545	50.7286	10.0000	10.0000	10.0000	51.7710	51.3856	52.1763	9.0000			0.9557	0.9298	0.7302	0.7761			0.5647	0.5707	
O	mlp	Rich leak-free text, names masked, Qwen3-Emb-8B	text_refseq_nomenfree	text		True	19256	0.0000	47.8778	47.1117	48.5910	11.0000	11.0000	12.0000	53.5145	53.0960	53.9289	6.0000			0.9461	0.9428	0.6975	0.7732			0.5746	0.5867	
O	mlp	GenePT ada-002, NCBI text as released	genept_ada	text		True	18942	0.3089	47.3128	46.5801	48.0026	12.0000	11.0000	13.0000	47.8962	47.4908	48.3126	15.0000	55.7564	56.7184	0.9536	0.9285	0.6943	0.7246	0.9054	0.9159	0.5618	0.5653	
O	mlp	GB.RNA-1.6B	rna1b6	rna		False	19228		47.2679	46.6028	47.9276	13.0000	11.0000	13.0000	45.9516	45.5084	46.3435	16.0000	54.0984	52.9526	0.8870	0.8673	0.6945	0.7091	0.8729	0.8698	0.6275	0.6129	
O	mlp	Geneformer V2-316M	geneformer	single-cell		True	19229		45.9383	45.2707	46.5839	14.0000	14.0000	15.0000	45.3732	44.9836	45.7705	18.0000	52.1585	51.3724	0.8636	0.8426	0.7375	0.7484	0.8541	0.8468	0.5879	0.5896	
O	mlp	Gene symbol only, Qwen3-Emb-8B	text_symbol	text		True	19256	0.0000	45.8568	45.1197	46.5711	15.0000	14.0000	15.0000	48.5508	48.1069	48.9534	14.0000			0.9330	0.9178	0.6980	0.7448			0.5569	0.5657	
O	mlp	GB.Cell-100M	cell_base	single-cell		True	18593		44.8516	44.2412	45.4993	16.0000	16.0000	17.0000	45.6076	45.2120	45.9936	17.0000	51.5046	51.6029	0.8659	0.8537	0.7270	0.7523	0.8573	0.8479	0.5798	0.5781	
O	mlp	scELMo, GPT-3.5-written descriptions	scelmo_gpt35	text		True	19231	0.0000	44.7627	44.0172	45.5147	17.0000	16.0000	17.0000	48.5784	48.1535	48.9845	13.0000	53.5631	57.6094	0.9322	0.9261	0.6802	0.7308	0.8998	0.9235	0.5591	0.5718	
O	mlp	GenePT text, generated summaries removed, Qwen3-Emb-8B	t23l__qwen8b	text		True	19256	0.0000	43.9468	43.2190	44.7528	18.0000	17.0000	18.0000	48.6372	48.2212	49.0615	12.0000			0.9245	0.9202	0.6786	0.7562			0.5561	0.5531	
O	mlp	Qwen2.5-7B-written descriptions, Qwen3-Emb-8B	llm_qwen25	text		True	19256	0.0000	40.3611	39.5890	41.0900	19.0000	19.0000	19.0000	44.0407	43.5920	44.4776	19.0000			0.9063	0.8932	0.6498	0.7100			0.5493	0.5574	
O	mlp	GB.DNA-300M	dna300m	dna		False	19236		38.9219	38.2901	39.5741	20.0000	20.0000	20.0000	36.3308	35.8974	36.7417	20.0000	44.4618	41.5327	0.8409	0.8136	0.6606	0.6603	0.8054	0.7857	0.5823	0.5711	
F	mlp	Qwen3-Embedding-8B	t23l__qwen8b	text		True	19256	0.0000	43.9468						48.6372						0.9245	0.9202	0.6786	0.7562			0.5561	0.5531	0.0334
F	mlp	pubmedbert-base-embeddings	t23l__pubmedbert	text		True	19256	0.0000	41.9467						46.1936						0.9235	0.9213	0.6490	0.7207			0.5567	0.5510	0.0404
F	mlp	all-mpnet-base-v2	t23l__mpnet	text		True	19256	0.0000	41.0171						44.9514						0.9207	0.9155	0.6398	0.7080			0.5548	0.5508	0.0382
F	mlp	Qwen3-Embedding-0.6B	t23l__qwen06b	text		True	19256	0.0000	41.0100						44.3086						0.9199	0.9122	0.6437	0.7079			0.5515	0.5445	0.0379
F	mlp	bge-large-en-v1.5	t23l__bge	text		True	19256	0.0000	41.0013						45.5100						0.9196	0.9150	0.6429	0.7170			0.5525	0.5506	0.0402
F	mlp	TF-IDF (1,024-d SVD)	t23l__tfidf	text		True	19256	0.0000	39.3258						39.4422						0.9202	0.9011	0.6252	0.6583			0.5445	0.5323	0.0210
