Determiner Productivity Evaluation on Manchester child dyads (test)

99.9BiasBabyBERTa-2

Evaluation results

Method
BabyBERTa-22026.05	99.9	0.005	0.005	49.7	0.445	0.001	58
BabyBERTa-32026.05	99.9	0.006	0.006	49.7	0.863	0.001	58
BabyBERTa-12026.05	99.6	0.014	0.014	49.7	0.057	0.001	58
babylm-baseline-10m-gpt-bert-mixed2026.05	98.4	0.052	0.049	48.7	0.128	0.001	58.2
babylm-baseline-100m-gpt-bert-masked-focus2026.05	98	0.062	0.054	48.9	0.043	0.001	58.2
babylm-baseline-100m-gpt-bert-mixed2026.05	97.6	0.074	0.067	48.5	0.013	0.001	58.1
babylm-baseline-10m-gpt-bert-masked-focus2026.05	95.7	0.118	0.101	48.8	0.003	0.001	58.1
ELC_BERT_small_baby_10M2026.05	94.9	0.143	0.116	49.8	0.001	0.001	58
gpt-bert-babylm-base2026.05	91.6	0.175	0.168	44.1	0.56	0.001	60.6
gpt-bert-babylm-small2026.05	90.2	0.185	0.185	43.1	0.997	0.001	62.6
roberta-base-10M-12026.05	86.2	0.225	0.227	36.7	0.907	0.001	70.2
roberta-base-10M-32026.05	83.6	0.267	0.25	34.7	0.296	0.001	72.9
Children2026.05	83.4	0.251	0.242	22.6	0.527	0.484	—
roberta-base-strict-20232026.05	83.4	0.301	0.249	47.5	0.001	0.001	58.5
roberta-base-10M-22026.05	82.3	0.278	0.26	33.8	0.328	0.001	74.6
Caretakers2026.05	81.5	0.301	0.32	20	0.191	0.256	—
opt-125m-strict-20232026.05	81.2	0.299	0.263	43	0.001	0.001	50
ELC_BERT_baby_100M2026.05	79.9	0.296	0.277	29.9	0.308	0.001	78.1
ltg-bert-babylm2026.05	79.7	0.297	0.278	31.4	0.289	0.001	77.4
ltgbert-10m-20242026.05	79.2	0.287	0.28	43.3	0.714	0.001	64
roberta-med-small-1M-32026.05	78.7	0.3	0.284	47.6	0.363	0.001	59.4
t5-base-strict-20232026.05	78.6	0.315	0.286	36	0.106	0.001	74.8
ltg-bert-bnc2026.05	78.3	0.314	0.287	22.5	0.138	0.505	86.6
babylm-baseline-100m-gpt-bert-causal-focus2026.05	78.2	0.314	0.288	24.7	0.122	0.041	89.8
babylm-baseline-100m-gpt-bert-mixed2026.05	77.7	0.319	0.291	25.1	0.113	0.023	88.9
roberta-base-strict-small-20232026.05	77.7	0.355	0.29	48.7	0.002	0.001	58.4
roberta-base2026.05	77.6	0.323	0.292	23.9	0.096	0.142	85.3
babyllama-100m-20242026.05	77.5	0.319	0.292	25.6	0.126	0.017	88.5
roberta-base-1B-32026.05	77.5	0.324	0.292	28.5	0.101	0.002	80.5
roberta-base-1B-12026.05	77	0.331	0.295	27.8	0.072	0.002	78.6
babyllama-10m-20242026.05	76.8	0.327	0.296	25.7	0.094	0.012	87
roberta-large2026.05	76.6	0.331	0.297	27	0.111	0.005	83.2
babylm-baseline-10m-gpt-bert-causal-focus2026.05	76.2	0.334	0.3	26.7	0.072	0.003	86.1
babylm-baseline-10m-gpt-bert-mixed2026.05	75.6	0.34	0.303	26.6	0.06	0.002	85.6
opt-125m2026.05	75.6	0.346	0.303	25.1	0.027	0.018	84.5
roberta-base-100M-12026.05	75.6	0.343	0.303	26.7	0.052	0.004	78.7
roberta-med-small-1M-12026.05	75.4	0.331	0.303	47.5	0.149	0.001	60.1
t5-base2026.05	75.3	0.347	0.304	23.9	0.03	0.085	84.6
gpt22026.05	75.2	0.349	0.305	25.2	0.031	0.012	83.5
roberta-base-100M-22026.05	75.2	0.343	0.305	28.5	0.07	0.001	79.6
roberta-base-100M-32026.05	74.8	0.353	0.308	28.2	0.03	0.001	78
roberta-base-1B-22026.05	74.7	0.351	0.308	28.3	0.031	0.001	79.7
opt-125m-strict-small-20232026.05	74.5	0.354	0.308	36.8	0.014	0.001	69.7
baby-llama-58m2026.05	74.1	0.352	0.309	27.1	0.025	0.002	81.7
babylm-baseline-10m-gpt22026.05	74	0.348	0.311	38.9	0.066	0.001	72.5
ltgbert-100m-20242026.05	73.9	0.352	0.311	30.8	0.055	0.001	77.6
babylm-baseline-100m-gpt22026.05	72.9	0.366	0.317	40.5	0.013	0.001	72.6
BabyLM-2026-Baseline-GPT2-Strict-Small2026.05	71.8	0.375	0.321	37.9	0.01	0.001	70
BabyLM-2026-Baseline-GPT2-Strict2026.05	70.1	0.381	0.327	39.3	0.007	0.001	65.3
roberta-med-small-1M-22026.05	69.3	0.39	0.332	45.9	0.01	0.001	58.7
t5-base-strict-small-20232026.05	65.7	0.42	0.341	48.3	0.001	0.001	43.6