mlsanityData to 14 Oct 2026

Mitigating Fabrication in Multi-Stage LLM Pipelines for Hiring: An Empirical Evaluation of Prompt Guardrails and Human-in-the-Loop Checkpoints

arXiv 2608.26171 ↗First posted 22 Jul 2026cs.CY cs.CL cs.HCcited by 1

Not on the map. Why?

Abstract (arXiv)

Multi-stage LLM hiring pipelines (resume improvement, interview question generation, answer feedback) can fabricate credentials, inflate qualifiers, and invent experience. We evaluate two mitigations, prompt guardrails and human-in-the-loop (HITL) checkpoints, against a fully automated baseline. In a controlled experiment (10 synthetic resumes x 2 job descriptions x 3 repetitions x 3 conditions; 180 runs), the baseline (C1) produced at least one unsupported claim in 96.7% of outputs (mean 6.80 findings/output). Prompt guardrails (C2) reduced finding density by 86% (6.80 to 0.92/output), but 50.0% of outputs still contained a fabrication, showing prompt-level mitigation alone is insufficient. A human checkpoint after resume improvement (C3) eliminated all identity fabrications, reduced finding density by 59% (6.88 to 2.82/output), reduced item-level fabrication from 96.7% to 75.0% (p=.022), and cut capture of JD-embedded trap requirements from 47% to 2% (vs. 5% under the guardrail). An exploratory analysis of multi-specialty resumes shows contamination rising monotonically with domain distance between specialties, suggesting career changers are especially exposed. The reviewer in this study caught all flagrant fabrications, but subtle qualifier drops and plausible new claims survived review roughly half the time (54.5% removal). Neither mitigation degraded the deliverable: claim retention exceeded 99% under both. The interventions are complementary: the guardrail eliminates unprompted additions and qualifier inflation cheaply, while the checkpoint gives near-categorical guarantees against the most severe failures, invented identities and JD-baited claims. These results support a layered architecture combining guardrails with a human checkpoint. A supplementary run with a newer-generation model (90.0% baseline fabrication rate) suggests the problem is not resolved by model progress alone.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 13 of its 23 references that are in the map. First posted on arXiv 22 Jul 2026.

Builds on

13 of 23 references.

References · 13

  1. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,601
  2. 2023Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversationscited by 1,207
  3. 2023Siren's Song in the AI Ocean: A Survey on Hallucination in Large Language Modelscited by 698
  4. 2023NeMo Guardrails: A Toolkit for Controllable and Safe LLM Applications with Programmable Railscited by 365
  5. 2024Self-Preference Bias in LLM-as-a-Judgecited by 258
  6. 2025Large Language Models Hallucination: A Comprehensive Surveycited by 49
  7. 2026From Spark to Fire: Modeling and Mitigating Error Cascades in LLM-Based Multi-Agent Collaborationcited by 14
  8. 2026Hallucination Cascade: Analyzing Error Propagation in Multi-Agent LLM Systemscited by 12
  9. 2026Self-Preference Bias in Rubric-Based Evaluation of Large Language Modelscited by 11
  10. 2026Quantifying and Mitigating Self-Preference Bias of LLM Judgescited by 10
  11. 2026PASC: Pipeline-Aware Conformal Prediction with Joint Coverage Guarantees for Multi-Stage NLP and LLM Pipelinescited by 7
  12. 2024ResumeFlow: An LLM-facilitated Pipeline for Personalized Resume Generation and Refinementcited by 6
1 more
  1. 2026(Human) Attention Is (Still) All You Need: Human oversight makes AI-assisted social science reliablecited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

Similar, not cited

Close in wording, no citation either way

  1. 2026Reviewer Capability Governs Rejection Targeting, Not Repair Skill: Evidence from LLM Execute-Review-Revise Pipelines
  2. 2025Let Them Down Easy! Contextual Effects of LLM Guardrails on User Perceptions and Preferences
  3. 2026Measuring Real-World Prompt Injection Attacks in LLM-based Resume Screening
  4. 2025ProCut: LLM Prompt Compression via Attribution Estimation
  5. 2026Prompt Injection in Automated R'esum'e Screening with Large Language Models: Single and Multi-Injection Settings
  6. 2025"You Are Rejected!": An Empirical Study of Large Language Models Taking Hiring Evaluations
  7. 2026Evaluating Prompt Injection Defenses for Educational LLM Tutors: Security-Usability-Latency Trade-offs
  8. 2024Assessing the Performance of Human-Capable LLMs -- Are LLMs Coming for Your Job?