AI Paper Atlas

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

STRIDE: Post-Training LLMs to Reason and Refine Bio-Sequences via Edit Trajectories

Daiheng Zhang, Shiyang Zhang, Sizhuang He, Yangtian Zhang, Syed A Rizvi, David van Dijk

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
add edit-trajectory supervision (full STRIDE) over direct final-sequence generation (ablation)Qwen3-14B · GFP replace-onlynovelty: 23 -> 53Improved
add structured edit tokens without rationale over direct generation (ablation)Qwen3-14B · GFP replace-onlynovelty: 23 -> 40Improved
add free-form rationale on top of structured edit tokens (ablation)STRIDE · GFP replace-onlysuccess: 48 -> 61Improved
use STRIDE instead of vanilla SFT, full action spaceQwen3-14B · GFP full actionsuccess (of 100): 42 -> 89Improved
use STRIDE instead of vanilla SFT, replace-onlyQwen3-14B · GFP replace-onlyunique improved: 39 -> 59Improved
use STRIDE instead of Edit Flow on AAVEdit Flow · AAV full actionsuccess: 52 -> 73Improved
add GSPO post-training to STRIDE-SFT (ablation)STRIDE-SFT · DrugAssiststrict success: 0.579 -> 0.676Improved
add GSPO post-training to STRIDE-SFT (ablation)STRIDE-SFT · DrugAssistshift rate: 0.983 -> 0.755Improved
add edit trajectories to GSPO RL (ablation)Vanilla GSPO · DrugAssistshift rate: 0.865 -> 0.755Improved
add explicit trajectory supervision in SFT without RL (ablation)Vanilla SFT · DrugAssiststrict success: 0.629 -> 0.579Hurt
use GRPO or CISPO instead of GSPO for STRIDE RL (ablation)STRIDE · DrugAssistshift rate: 0.755 -> 0.956Hurt
closed-loop 3x30 refinement instead of one-shot 1x90STRIDE · GFP replace-onlynovelty: 84.3% -> 97.0%Improved

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