AI Paper Atlas

Research map / Large language models / Instruction and fine-tuning

Do We Need Adam? Surprisingly Strong and Sparse Reinforcement Learning with SGD in LLMs

Sagnik Mukherjee, Lifan Yuan, Pavan Jayasinha, Dilek Hakkani-Tür, Hao Peng

ICML 2026 · Large language models › Instruction and fine-tuning · approach: Fine-tuning for reasoning data · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace AdamW with SGDQwen3-8B GRPO · Math mean (3K)mean accuracy: 55.6 -> 55.8No clear effect
replace AdamW with SGDQwen3-1.7B GRPO · Math mean (3K)mean accuracy: 46.0 -> 48.6Improved
replace AdamW with SGDQwen3-1.7B GRPO · Math mean (8K)mean accuracy: 58.2 -> 56.8Hurt
replace AdamW with SGDQwen3-1.7B GRPO · HumanEval/MBPP pass@1 (4K)pass@1: +8.7 average pass@1Improved
replace AdamW with SGDQwen3-1.7B GRPO · MBPP+ pass@1 (4K)pass@1: 46.8 -> 59.0Improved
add momentum to SGD (ablation)Qwen3-8B GRPO · Math mean (3K)mean accuracy: 55.8 -> 47.1Hurt
add momentum to SGD (ablation)Qwen3-1.7B GRPO · Math mean (3K)mean accuracy: 48.6 -> 48.9Improved
add adaptive learning rate (RMSProp) to SGD (ablation)Qwen3-8B GRPO · Math mean (8K)mean accuracy: 70.0 -> 71.1Improved
add adaptive learning rate (RMSProp) to SGD (ablation)Llama-3.1-8B GRPO · Math mean (3K)mean accuracy: 29.4 -> 29.5No clear effect
replace AdamW with SGDQwen3-8B PPO · Math meanmean accuracy: 54.0 -> 56.7Improved
replace AdamW with SGDQwen3-1.7B GRPO on RLVE · Math mean (16K)mean accuracy: 64.5 -> 63.0No clear effect
replace AdamW with SGDQwen3-8B math GRPO · parameter update sparsity (%)sparsity: 91.30 -> 99.99Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy