router-expert-math_word

LoRA expert for the math_word target of the route-then-admit pool, trained on meirdick/router-experts-data data/experts/math_word.jsonl.

• base: Qwen/Qwen3-4B-Instruct-2507
• rank 16, alpha 32, dropout 0.05, modules q_proj, k_proj, v_proj, o_proj
• lr 0.0002, epochs 2, max_len 1024, token budget 8192 per batch
• one example per item per recipe (direct,cot_short), under the serving system prompt; 40% of the mc items re-lettered to 5 to 10 options
• loss on the assistant turn only; held-out 5% for the val loss

field	value
target	math_word
items	1500
recipes	direct,cot_short
kinds	{'numeric': 1500}
mc_padded	0
examples	3000
encoded	3000
dropped_too_long	0
train_rows	2850
val_rows	150
train_batches	36
train_tokens	283425
supervised_tokens	98126
steps	72
train_loss	0.5862834812659357
val_loss_before	2.9753416776416812
val_loss	0.2892059890932192
seconds	130.87
trainable_params	11796480