Length-Normalized Advantage Formulation

{
  "doc": "nguyen-2025-sfr-deepresearch.md",
  "context": "Novel modification to REINFORCE that normalizes step-level advantage by trajectory length to prevent long but low-quality trajectories from dominating training.",
  "category": "ai",
  "norm_label": "Length-Normalized Advantage Formulation",
  "graphify_id": "length_normalized_advantage",
  "source_file": "nguyen-2025-sfr-deepresearch.md",
  "imported_from": "/Users/antonborzov/Documents/Research.nosync/papers/extract_typed_out/nguyen-2025-sfr-deepresearch/graph.json",
  "extracted_type": "method",
  "source_location": "§3.3, Eq. 1",
  "graphify_file_type": "method"
}