Modelpublic

lsteno/Qwen3-4B-Instruct-2507-RLM-RLVR-FullFT-lr5e-6-depth1-v1

sourceHugging Faceapache-2.0updated 4mo agoView on Hugging Face

0likes20downloads

Model Card

Full-parameter RLM RLVR checkpoint.

—Base model: Qwen/Qwen3-4B-Instruct-2507
—Source checkpoint: /home/ubuntu/prime_run/outputs/rlm-rlvr-qwen3-4b-depth1-llmonly-fullft-lr5e-6-s150-bal35f40v1-ncclfixed-allrollouts/weights/step_150
—Step: 150
—Prompt variant: sanjaya_text_depth1_llm_only_v1
—Runtime: depth-1 LLM-only RLM harness, plain Gemini subcalls, recursive child RLMs disabled.