Skip to content
← Back to explorer

Controllable Reasoning Models Are Private Thinkers

Haritz Puerto, Haonan Li, Xudong Han, Timothy Baldwin, Iryna Gurevych · Feb 27, 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

Feb 27, 2026, 5:39 PM

Recent

Extraction refreshed

Mar 8, 2026, 6:21 AM

Fresh

Extraction source

Runtime deterministic fallback

Confidence 0.35

Abstract

AI agents powered by reasoning models require access to sensitive user data. However, their reasoning traces are difficult to control, which can result in the unintended leakage of private information to external parties. We propose training models to follow instructions not only in the final answer, but also in reasoning traces, potentially under different constraints. We hypothesize that improving their instruction following abilities in the reasoning traces can improve their privacy-preservation skills. To demonstrate this, we fine-tune models on a new instruction-following dataset with explicit restrictions on reasoning traces. We further introduce a generation strategy that decouples reasoning and answer generation using separate LoRA adapters. We evaluate our approach on six models from two model families, ranging from 1.7B to 14B parameters, across two instruction-following benchmarks and two privacy benchmarks. Our method yields substantial improvements, achieving gains of up to 20.9 points in instruction-following performance and up to 51.9 percentage points on privacy benchmarks. These improvements, however, can come at the cost of task utility, due to the trade-off between reasoning performance and instruction-following abilities. Overall, our results show that improving instruction-following behavior in reasoning models can significantly enhance privacy, suggesting a promising direction for the development of future privacy-aware agents. Our code and data are available at https://github.com/UKPLab/arxiv2026-controllable-reasoning-models

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: Runtime deterministic fallback missing

No explicit feedback protocol extracted.

Evidence snippet: AI agents powered by reasoning models require access to sensitive user data.

Evaluation Modes

partial

Automatic Metrics

Confidence: Low Source: Runtime deterministic fallback evidenced

Includes extracted eval setup.

Evidence snippet: AI agents powered by reasoning models require access to sensitive user data.

Quality Controls

missing

Not reported

Confidence: Low Source: Runtime deterministic fallback missing

No explicit QC controls found.

Evidence snippet: AI agents powered by reasoning models require access to sensitive user data.

Benchmarks / Datasets

missing

Not extracted

Confidence: Low Source: Runtime deterministic fallback missing

No benchmark anchors detected.

Evidence snippet: AI agents powered by reasoning models require access to sensitive user data.

Reported Metrics

partial

Cost

Confidence: Low Source: Runtime deterministic fallback evidenced

Useful for evaluation criteria comparison.

Evidence snippet: These improvements, however, can come at the cost of task utility, due to the trade-off between reasoning performance and instruction-following abilities.

Rater Population

missing

Unknown

Confidence: Low Source: Runtime deterministic fallback missing

Rater source not explicitly reported.

Evidence snippet: AI agents powered by reasoning models require access to sensitive user data.

Human Data Lens

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

Evaluation Lens

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

Protocol And Measurement Signals

Benchmarks / Datasets

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

Reported Metrics

cost

Research Brief

Deterministic synthesis

AI agents powered by reasoning models require access to sensitive user data. HFEPX signals include Automatic Metrics with confidence 0.35. Updated from current HFEPX corpus.

Generated Mar 8, 2026, 6:21 AM · Grounded in abstract + metadata only

Key Takeaways

  • AI agents powered by reasoning models require access to sensitive user data.
  • We propose training models to follow instructions not only in the final answer, but also in reasoning traces, potentially under different constraints.

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 (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

  • AI agents powered by reasoning models require access to sensitive user data.
  • We propose training models to follow instructions not only in the final answer, but also in reasoning traces, potentially under different constraints.
  • We evaluate our approach on six models from two model families, ranging from 1.7B to 14B parameters, across two instruction-following benchmarks and two privacy benchmarks.

Why It Matters For Eval

  • AI agents powered by reasoning models require access to sensitive user data.
  • We evaluate our approach on six models from two model families, ranging from 1.7B to 14B parameters, across two instruction-following benchmarks and two privacy benchmarks.

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: 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