List your company
ResearchFormerly WizWand

Symbolic Regression on LSR-Synth Physics

0.0001NMSEPiT-PO
-0.0061320.0357010.0775330.119365Feb 11, 2026Mar 1, 2026Mar 20, 2026Apr 8, 2026Apr 26, 2026May 15, 2026Jun 3, 2026
Updated 5mo ago

Evaluation results

MethodLinks
PiT-POBackbone=Llama-3.1-8B-Instruct2026.020.000111.3640.91—
LLM-SR+Backbone=Qwen2.5-7B, Optimization Strategy=SAGE-Fit (Inner-loop)2026.050.000111.3648—
LLM-SRBackbone=Qwen2.5-7B, Optimization Strategy=Post-hoc Refinement2026.050.00014.5536.84—
LLM-SRBackbone=Llama-3.1-8B-Instruct2026.020.00016.9734.09—
Deliberate EvolutionBackbone=Qwen3-4B-Instruct-25072026.060.0004——15.91
Deliberate EvolutionBackbone=Llama3.1-8B-Instruct2026.060.001——11.36
LaSRBackbone=Llama-3.1-8B-Instruct2026.020.00184.5425.02—
LLM-SRBackbone=Qwen3-4B-Instruct-25072026.060.0025——6.82
PySR+Optimization Strategy=SAGE-Fit (Inner-loop)2026.050.00292.2752.63—
LLM-SRBackbone=Llama3.1-8B-Instruct2026.060.003——6.82
LaSR+Backbone=Qwen2.5-7B, Optimization Strategy=SAGE-Fit (Inner-loop)2026.050.00396.819.09—
PySROptimization Strategy=Post-hoc Refinement2026.050.0048016—
uDSR+Optimization Strategy=SAGE-Fit (Inner-loop)2026.050.00482.2712.5—
LaSRBackbone=Qwen2.5-7B, Optimization Strategy=Post-hoc Refinement2026.050.0062.276.82—
LASRBackbone=Qwen3-4B-Instruct-25072026.060.006——6.82
LASRBackbone=Llama3.1-8B-Instruct2026.060.0061——9.09
LLMDirectBackbone=Llama3.1-8B-Instruct2026.060.01——0
uDSROptimization Strategy=Post-hoc Refinement2026.050.011907.5—
Direct PromptingBackbone=Llama-3.1-8B-Instruct2026.020.045900—
LLMDirectBackbone=Qwen3-4B-Instruct-25072026.060.0546——6.82
SGABackbone=Qwen3-4B-Instruct-25072026.060.104——0
SGABackbone=Llama-3.1-8B-Instruct2026.020.154902.27—
SGABackbone=Llama3.1-8B-Instruct2026.060.155——2.27