CITATION.cff

cff-version: 1.2.0
message: "If you use this software or reference this work, please cite it as below."
title: "Branch-Aware Hindsight Credit Assignment for Language-Model Agents Under Matched Budgets"
authors:
  - family-names: Fix
    given-names: Jaden
type: software
license: MIT
repository-code: "https://github.com/jadenfix/CreditAssignment"
keywords:
  - credit assignment
  - language model agents
  - process reward models
  - reinforcement learning
  - ALFWorld
abstract: >-
  We identify loop-action bias as a mechanistic failure mode of outcome-only
  credit assignment for LLM agents and propose branch-aware and hindsight
  scorers that leverage per-transition verifier signals. Validated on controlled
  stochastic benchmarks and the real ALFWorld environment with two LLMs using
  automatically extracted verifier signals.