Modelpublic

swapnil7777/sfpo-sfpo-llama-iso-3b-iso-shadowk-5-shutoff-adaptive-hendrycks-math-seed42-20260414-06-8cb42bc6

sourceHugging Faceupdated 6mo agoView on Hugging Face

0likes5downloads

Model Card

This repo was uploaded from a local training checkpoint.

—Source run: sfpo_llama_iso_3B_iso_shadowk_5_shutoff_adaptive_hendrycks_math_seed42_20260414_060012
—Checkpoint: checkpoint-224
—Local path: /home/ismam/lookahead/lookahead_codes/checkpoints_hendrycks_math/sfpo_llama_iso_3B_iso_shadowk_5_shutoff_adaptive_hendrycks_math_seed42_20260414_060012/checkpoint-224
—Collection: swapnil7777/sfpo-math-llama-69de7b59096a251ce8af5720
—Generated at: 2026-04-15 09:27:37 UTC

This checkpoint name is designed to stay readable:

—run name preserves method, model, K, dataset, seed, and run timestamp
—checkpoint suffix preserves whether this is best_checkpoint, bp_budget_*, or checkpoint-*