List your company
ResearchFormerly WizWand

LLM Similarity Measurement on RL Models

100Similarity ScoreLogits
43.8458.427387.58Jul 3, 2026
Updated 4mo ago

Evaluation results

MethodLinks
LogitsRL Model/Algorithm=Archer-Code-1.5B (GRPO)2026.07100
GhostSpecRL Model/Algorithm=Archer-Code-1.5B (GRPO)2026.07100
LogitsRL Model/Algorithm=Nemotron-Research-Reasoning-Qwen-1.5B (Reinforcement++)2026.07100
GhostSpecRL Model/Algorithm=Nemotron-Research-Reasoning-Qwen-1.5B (Reinforcement++)2026.07100
PCSRL Model/Algorithm=Polaris-7B-Preview (DAPO)2026.07100
LogitsRL Model/Algorithm=Polaris-7B-Preview (DAPO)2026.07100
PCSRL Model/Algorithm=Archer-Code-1.5B (GRPO)2026.0799
Spectral SignatureRL Model/Algorithm=Archer-Code-1.5B (GRPO)2026.0799
PCSRL Model/Algorithm=Nemotron-Research-Reasoning-Qwen-1.5B (Reinforcement++)2026.0799
Spectral SignatureRL Model/Algorithm=Nemotron-Research-Reasoning-Qwen-1.5B (Reinforcement++)2026.0799
Spectral SignatureRL Model/Algorithm=Polaris-7B-Preview (DAPO)2026.0799
REEFRL Model/Algorithm=Archer-Code-1.5B (GRPO)2026.0798
REEFRL Model/Algorithm=Nemotron-Research-Reasoning-Qwen-1.5B (Reinforcement++)2026.0796
REEFRL Model/Algorithm=Polaris-7B-Preview (DAPO)2026.0794
GhostSpecRL Model/Algorithm=Polaris-7B-Preview (DAPO)2026.0746