Skip to content
OpenTrain AIFor AI Companies

HFEPX · Eval paper review

DistillGuard: Evaluating Defenses Against LLM Knowledge Distillation

Bo Jiang

Published

Mar 8, 2026

Citations

0

Trust level

Low

Usefulness score

0/100 (Low)

Extraction confidence

25% (Low)

Derived from extracted protocol signals and abstract evidence.

Rater population

Not reported

Signals refreshed

Mar 8, 2026

Open arXiv Find implementation Shortlist (0) RSS feed

Should you rely on this paper?

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

Use this as background context only. Do not make protocol decisions from this page alone.

Best use

Background context only

Use if you need

Background context only.

What to verify

Validate the evaluation procedure and quality controls in the full paper before operational use.

Main weakness

This paper looks adjacent to evaluation work, but not like a strong protocol reference.

Human feedback signal
Not explicit
Not explicit in abstract metadata
Evaluation signal
Weak or implicit
Validate from full paper
Usefulness for eval research
0/100
Adjacent candidate

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

If you are doing eval pipeline work, start here

Human Eval Papers LLM-as-Judge Papers Pairwise Preference Papers Tool-Use Evaluation Search

Abstract

Knowledge distillation from proprietary LLM APIs poses a growing threat to model providers, yet defenses against this attack remain fragmented and unevaluated. We present DistillGuard, a framework for systematically evaluating output-level defenses against LLM knowledge distillation. We introduce a taxonomy of three defense categories -- output perturbation, data poisoning, and information throttling -- and evaluate nine defense configurations using a standardized pipeline with Qwen3-14B as teacher and Qwen2.5-7B-Instruct as student across three benchmarks (MATH-500, HumanEval+, MT-Bench). Our results reveal that, in a same-family distillation setting against a naive attacker, most output-level defenses are surprisingly ineffective: paraphrasing-based perturbation barely degrades distilled student quality, and data poisoning primarily impairs conversational fluency while leaving task-specific capabilities intact. Only chain-of-thought removal substantially impairs mathematical reasoning (31.4\% vs.\ 67.8\% baseline), though code generation remains unaffected. These findings demonstrate that the effectiveness of distillation defenses is highly task-dependent and that current output-level approaches are insufficient to broadly prevent knowledge theft.

What we could verify

These are the protocol signals we could actually recover from the available paper metadata. Use them to decide whether this paper is worth deeper reading.

Human Feedback Types

missing

None explicit

No explicit feedback protocol extracted.

"Knowledge distillation from proprietary LLM APIs poses a growing threat to model providers, yet defenses against this attack remain fragmented and unevaluated."

Evaluation Modes

missing

None explicit

Validate eval design from full paper text.

"Knowledge distillation from proprietary LLM APIs poses a growing threat to model providers, yet defenses against this attack remain fragmented and unevaluated."

Quality Controls

missing

Not reported

No explicit QC controls found.

"Knowledge distillation from proprietary LLM APIs poses a growing threat to model providers, yet defenses against this attack remain fragmented and unevaluated."

Benchmarks / Datasets

partial

MT Bench, MATH 500, HumanEval+

Useful for quick benchmark comparison.

"We introduce a taxonomy of three defense categories -- output perturbation, data poisoning, and information throttling -- and evaluate nine defense configurations using a standardized pipeline with Qwen3-14B as teacher and Qwen2.5-7B-Instruct as student across three benchmarks (MATH-500, HumanEval+, MT-Bench)."

Reported Metrics

missing

Not extracted

No metric anchors detected.

"Knowledge distillation from proprietary LLM APIs poses a growing threat to model providers, yet defenses against this attack remain fragmented and unevaluated."

Benchmarks and datasets

MT-BenchMATH-500HumanEval+

Reported metrics

No metric terms were extracted from the available abstract.

Human feedback details
Uses human feedback
No
Feedback types
None
Rater population
Not reported
Expertise required
Math, Coding
Evaluation details
Evaluation modes
None
Agentic eval
None
Quality controls
Not reported
Evidence quality
Low
Use this page as
Background context only

Research brief

Metadata summary

Knowledge distillation from proprietary LLM APIs poses a growing threat to model providers, yet defenses against this attack remain fragmented and unevaluated.

Based on abstract + metadata only. Check the source paper before making high-confidence protocol decisions.

Key takeaways

  • Knowledge distillation from proprietary LLM APIs poses a growing threat to model providers, yet defenses against this attack remain fragmented and unevaluated.
  • We present DistillGuard, a framework for systematically evaluating output-level defenses against LLM knowledge distillation.
  • We introduce a taxonomy of three defense categories -- output perturbation, data poisoning, and information throttling -- and evaluate nine defense configurations using a standardized pipeline with Qwen3-14B as teacher and Qwen2.5-7B-Instruct as student across three benchmarks (MATH-500, HumanEval+, MT-Bench).

Researcher actions

  • Compare this paper against others mentioning MT-Bench.
  • Check the full text for explicit evaluation design choices (raters, protocol, and metrics).
  • Use related-paper links to find stronger protocol-specific references.

Caveats

  • Generated from abstract + metadata only; no PDF parsing.
  • Signals below are heuristic and may miss details reported outside the abstract.

Recommended queries

MT-Bench

Contribution summary

  • We present DistillGuard, a framework for systematically evaluating output-level defenses against LLM knowledge distillation.
  • We introduce a taxonomy of three defense categories -- output perturbation, data poisoning, and information throttling -- and evaluate nine defense configurations using a standardized pipeline with Qwen3-14B as teacher and…

Why it matters for eval

  • We introduce a taxonomy of three defense categories -- output perturbation, data poisoning, and information throttling -- and evaluate nine defense configurations using a standardized pipeline with Qwen3-14B as teacher and…

Researcher checklist

  • Human feedback protocol is explicit

    No explicit human feedback protocol detected.

  • Evaluation mode is explicit

    No clear evaluation mode extracted.

  • Quality control reporting appears

    No calibration/adjudication/IAA control explicitly detected.

  • Benchmark or dataset anchors are present

    Detected: MT-Bench, MATH-500, HumanEval+

  • Metric reporting is present

    No metric terms extracted.