Skip to content
← Back to explorer

FrugalRAG: Less is More in RL Finetuning for Multi-Hop Question Answering

Abhinav Java, Srivathsan Koundinyan, Nagarajan Natarajan, Amit Sharma · Jul 10, 2025 · Citations: 0

Automatic Metrics
Open arXiv Find Implementation RSS feed Shortlist (0)

Data freshness

Extraction: Fresh

Check recency before relying on this page for active eval decisions. Use stale pages as context and verify against current hub results.

Metadata refreshed

Feb 28, 2026, 12:24 AM

Recent

Extraction refreshed

Mar 7, 2026, 9:02 PM

Fresh

Extraction source

Runtime deterministic fallback

Confidence 0.45

Abstract

Reinforcement learning (RL) based on the final answer's reward has driven recent progress in small language models (SLMs) on reasoning-heavy tasks such as math and code. However, applying the same techniques to retrieval-augmented generation (RAG) benchmarks like multi-hop QA has yielded limited gains, often trailing supervised or prompting-only baselines. Instead, we argue that a viable path for RL in multi-hop QA is to use test-time scaling judiciously to optimize both final answer accuracy and efficiency in reaching that answer. We propose FrugalRAG, a two-stage finetuning framework that adaptively reduces the number of retrieval steps based on a question's difficulty. First, we train an SLM with supervised finetuning on a full-exploration policy that generates broad sub-queries. Then, we apply RL to adaptively prune search depth based on question difficulty, directly rewarding policies that balance correctness with frugality. Unlike prior approaches requiring 10x more data, our method achieves competitive performance with only approximately 1,000 examples. On HotPotQA and other multi-hop QA benchmarks, FrugalRAG attains state-of-the-art efficiency-accuracy tradeoffs, cutting retrieval cost nearly in half. Moreover, on the challenging BrowseCompPlus benchmark, it generalizes zero-shot and surpasses SLM-based and other baselines. These results demonstrate the use of RL not to increase reasoning steps, but to reduce them, as an effective solution for scalable and efficient RAG.

Low-signal caution for protocol decisions

Use this page for context, then validate protocol choices against stronger HFEPX references before implementation decisions.

  • Extraction flags indicate low-signal or possible false-positive protocol mapping.
  • Extraction confidence is 0.45 (below strong-reference threshold).
Human Eval Hub LLM-as-Judge Hub Pairwise Preference Hub

HFEPX Relevance Assessment

This paper is adjacent to HFEPX scope and is best used for background context, not as a primary protocol reference.

Best use

Background context only

Use if you need

A benchmark-and-metrics comparison anchor.

Main weakness

Extraction flags indicate low-signal or possible false-positive protocol mapping.

Trust level

Low

Eval-Fit Score

5/100 • Low

Treat as adjacent context, not a core eval-method reference.

Human Feedback Signal

Not explicit in abstract metadata

Evaluation Signal

Detected

HFEPX Fit

Adjacent candidate

Extraction confidence: Low

If you are doing eval pipeline work, start here:

Human Eval Hub LLM-as-Judge Hub Pairwise Preference Hub Tool-Use Eval Hub

Field Provenance & Confidence

Each key protocol field shows extraction state, confidence band, and data source so you can decide whether to trust it directly or validate from full text.

Human Feedback Types

missing

None explicit

Confidence: Low Source: Runtime deterministic fallback missing

No explicit feedback protocol extracted.

Evidence snippet: Reinforcement learning (RL) based on the final answer's reward has driven recent progress in small language models (SLMs) on reasoning-heavy tasks such as math and code.

Evaluation Modes

partial

Automatic Metrics

Confidence: Low Source: Runtime deterministic fallback evidenced

Includes extracted eval setup.

Evidence snippet: Reinforcement learning (RL) based on the final answer's reward has driven recent progress in small language models (SLMs) on reasoning-heavy tasks such as math and code.

Quality Controls

missing

Not reported

Confidence: Low Source: Runtime deterministic fallback missing

No explicit QC controls found.

Evidence snippet: Reinforcement learning (RL) based on the final answer's reward has driven recent progress in small language models (SLMs) on reasoning-heavy tasks such as math and code.

Benchmarks / Datasets

partial

HotpotQA

Confidence: Low Source: Runtime deterministic fallback evidenced

Useful for quick benchmark comparison.

Evidence snippet: On HotPotQA and other multi-hop QA benchmarks, FrugalRAG attains state-of-the-art efficiency-accuracy tradeoffs, cutting retrieval cost nearly in half.

Reported Metrics

partial

Accuracy, Cost

Confidence: Low Source: Runtime deterministic fallback evidenced

Useful for evaluation criteria comparison.

Evidence snippet: Instead, we argue that a viable path for RL in multi-hop QA is to use test-time scaling judiciously to optimize both final answer accuracy and efficiency in reaching that answer.

Rater Population

missing

Unknown

Confidence: Low Source: Runtime deterministic fallback missing

Rater source not explicitly reported.

Evidence snippet: Reinforcement learning (RL) based on the final answer's reward has driven recent progress in small language models (SLMs) on reasoning-heavy tasks such as math and code.

Human Data Lens

  • Uses human feedback: No
  • Feedback types: None
  • Rater population: Unknown
  • Unit of annotation: Unknown
  • Expertise required: Math, Coding
  • Extraction source: Runtime deterministic fallback

Evaluation Lens

  • Evaluation modes: Automatic Metrics
  • Agentic eval: None
  • Quality controls: Not reported
  • Confidence: 0.45
  • Flags: low_signal, possible_false_positive, runtime_fallback_extraction

Protocol And Measurement Signals

Benchmarks / Datasets

HotpotQA

Reported Metrics

accuracycost

Research Brief

Deterministic synthesis

However, applying the same techniques to retrieval-augmented generation (RAG) benchmarks like multi-hop QA has yielded limited gains, often trailing supervised or prompting-only baselines. HFEPX signals include Automatic Metrics with confidence 0.45. Updated from current HFEPX corpus.

Generated Mar 7, 2026, 9:02 PM · Grounded in abstract + metadata only

Key Takeaways

  • However, applying the same techniques to retrieval-augmented generation (RAG) benchmarks like multi-hop QA has yielded limited gains, often trailing supervised or prompting-only…
  • We propose FrugalRAG, a two-stage finetuning framework that adaptively reduces the number of retrieval steps based on a question's difficulty.

Researcher Actions

  • Treat this as method context, then pivot to protocol-specific HFEPX hubs.
  • Cross-check benchmark overlap: HotpotQA.
  • Validate metric comparability (accuracy, cost).

Caveats

  • Generated from title, abstract, and extracted metadata only; full-paper implementation details are not parsed.
  • Low-signal flag detected: protocol relevance may be indirect.

Research Summary

Contribution Summary

  • However, applying the same techniques to retrieval-augmented generation (RAG) benchmarks like multi-hop QA has yielded limited gains, often trailing supervised or prompting-only baselines.
  • We propose FrugalRAG, a two-stage finetuning framework that adaptively reduces the number of retrieval steps based on a question's difficulty.
  • On HotPotQA and other multi-hop QA benchmarks, FrugalRAG attains state-of-the-art efficiency-accuracy tradeoffs, cutting retrieval cost nearly in half.

Why It Matters For Eval

  • However, applying the same techniques to retrieval-augmented generation (RAG) benchmarks like multi-hop QA has yielded limited gains, often trailing supervised or prompting-only baselines.
  • On HotPotQA and other multi-hop QA benchmarks, FrugalRAG attains state-of-the-art efficiency-accuracy tradeoffs, cutting retrieval cost nearly in half.

Researcher Checklist

  • Gap: Human feedback protocol is explicit

    No explicit human feedback protocol detected.

  • Pass: Evaluation mode is explicit

    Detected: Automatic Metrics

  • Gap: Quality control reporting appears

    No calibration/adjudication/IAA control explicitly detected.

  • Pass: Benchmark or dataset anchors are present

    Detected: HotpotQA

  • Pass: Metric reporting is present

    Detected: accuracy, cost

Category-Adjacent Papers (Broader Context)

These papers are nearby in arXiv category and useful for broader context, but not necessarily protocol-matched to this paper.

Need human evaluators for your AI research? Scale annotation with expert AI Trainers.

Post a Job Get a Quote