Skip to content
← Back to explorer

Language-Aware Distillation for Multilingual Instruction-Following Speech LLMs with ASR-Only Supervision

Shreyas Gopal, Donghang Wu, Ashutosh Anshul, Yeo Yue Heng, Yizhou Peng, Haoyang Li, Hexin Liu, Eng Siong Chng · Mar 7, 2026 · 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

Mar 7, 2026, 4:09 AM

Recent

Extraction refreshed

Mar 14, 2026, 5:04 AM

Fresh

Extraction source

Persisted extraction

Confidence 0.35

Abstract

Speech Large Language Models (LLMs) that understand and follow instructions in many languages are useful for real-world interaction, but are difficult to train with supervised fine-tuning, requiring large, task-specific speech corpora. While recent distillation-based approaches train performant English-only Speech LLMs using only annotated ASR data by aligning text and speech using only a lightweight projector, these models under-perform when scaled to multilingual settings due to language interference in the shared projector. We address this by introducing language-aware distillation using a query bank and a gating network that selects or mixes query tokens using a Q-Former projector. Our approach shows gains of 14% over matched multilingual distillation baselines on instruction following. We further synthesize Audio-MLQA, a multilingual spoken QA benchmark built on MLQA with high-quality TTS questions. Our best model improves over existing Speech LLM baselines by 32% on Audio-MLQA.

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 0.35 (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 secondary eval reference to pair with stronger protocol papers.

Main weakness

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

Trust level

Low

Eval-Fit Score

0/100 • 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: Persisted extraction missing

No explicit feedback protocol extracted.

Evidence snippet: Speech Large Language Models (LLMs) that understand and follow instructions in many languages are useful for real-world interaction, but are difficult to train with supervised fine-tuning, requiring large, task-specific speech corpora.

Evaluation Modes

partial

Automatic Metrics

Confidence: Low Source: Persisted extraction evidenced

Includes extracted eval setup.

Evidence snippet: Speech Large Language Models (LLMs) that understand and follow instructions in many languages are useful for real-world interaction, but are difficult to train with supervised fine-tuning, requiring large, task-specific speech corpora.

Quality Controls

missing

Not reported

Confidence: Low Source: Persisted extraction missing

No explicit QC controls found.

Evidence snippet: Speech Large Language Models (LLMs) that understand and follow instructions in many languages are useful for real-world interaction, but are difficult to train with supervised fine-tuning, requiring large, task-specific speech corpora.

Benchmarks / Datasets

missing

Not extracted

Confidence: Low Source: Persisted extraction missing

No benchmark anchors detected.

Evidence snippet: Speech Large Language Models (LLMs) that understand and follow instructions in many languages are useful for real-world interaction, but are difficult to train with supervised fine-tuning, requiring large, task-specific speech corpora.

Reported Metrics

partial

Jailbreak success rate

Confidence: Low Source: Persisted extraction evidenced

Useful for evaluation criteria comparison.

Evidence snippet: Speech Large Language Models (LLMs) that understand and follow instructions in many languages are useful for real-world interaction, but are difficult to train with supervised fine-tuning, requiring large, task-specific speech corpora.

Rater Population

missing

Unknown

Confidence: Low Source: Persisted extraction missing

Rater source not explicitly reported.

Evidence snippet: Speech Large Language Models (LLMs) that understand and follow instructions in many languages are useful for real-world interaction, but are difficult to train with supervised fine-tuning, requiring large, task-specific speech corpora.

Human Data Lens

  • Uses human feedback: No
  • Feedback types: None
  • Rater population: Unknown
  • Unit of annotation: Unknown
  • Expertise required: Multilingual
  • Extraction source: Persisted extraction

Evaluation Lens

  • Evaluation modes: Automatic Metrics
  • Agentic eval: None
  • Quality controls: Not reported
  • Confidence: 0.35
  • Flags: low_signal, possible_false_positive

Protocol And Measurement Signals

Benchmarks / Datasets

No benchmark or dataset names were extracted from the available abstract.

Reported Metrics

jailbreak success rate

Research Brief

Deterministic synthesis

Our approach shows gains of 14% over matched multilingual distillation baselines on instruction following. HFEPX signals include Automatic Metrics with confidence 0.35. Updated from current HFEPX corpus.

Generated Mar 14, 2026, 5:04 AM · Grounded in abstract + metadata only

Key Takeaways

  • Our approach shows gains of 14% over matched multilingual distillation baselines on instruction following.
  • We further synthesize Audio-MLQA, a multilingual spoken QA benchmark built on MLQA with high-quality TTS questions.

Researcher Actions

  • Treat this as method context, then pivot to protocol-specific HFEPX hubs.
  • Identify benchmark choices from full text before operationalizing conclusions.
  • Validate metric comparability (jailbreak success rate).

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

  • Our approach shows gains of 14% over matched multilingual distillation baselines on instruction following.
  • We further synthesize Audio-MLQA, a multilingual spoken QA benchmark built on MLQA with high-quality TTS questions.
  • Our best model improves over existing Speech LLM baselines by 32% on Audio-MLQA.

Why It Matters For Eval

  • We further synthesize Audio-MLQA, a multilingual spoken QA benchmark built on MLQA with high-quality TTS questions.

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.

  • Gap: Benchmark or dataset anchors are present

    No benchmark/dataset anchor extracted from abstract.

  • Pass: Metric reporting is present

    Detected: jailbreak success rate

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