Skip to content
← Back to explorer

HFEPX Hub

CS.CR Human Feedback And Eval Papers

Updated from current HFEPX corpus (Mar 8, 2026). 57 papers are grouped in this hub page.

Read Full Context

Updated from current HFEPX corpus (Mar 8, 2026). 57 papers are grouped in this hub page. Common evaluation modes: Automatic Metrics, Llm As Judge. Most common rater population: Domain Experts. Common annotation unit: Trajectory. Frequent quality control: Calibration. Frequently cited benchmark: AdvBench. Common metric signal: jailbreak success rate. Use this page to compare protocol setup, judge behavior, and labeling design decisions before running new eval experiments. Newest paper in this set is from Feb 27, 2026.

Papers: 57 Last published: Feb 27, 2026 Global RSS
Cs.CR

Researcher Quick Triage

This hub is best used for protocol triage and replication planning from abstract-level evidence. Quality band: Medium .

All Sampled Papers (57) Replication-Ready Only (6)

High-Signal Coverage

100.0%

57 / 57 sampled papers are not low-signal flagged.

Replication-Ready Set

6

Benchmark + metric + eval mode explicitly present.

Judge/Human Comparability

0

Papers containing both `human_eval` and `llm_as_judge`.

  • 6 papers are replication-ready (benchmark + metric + explicit evaluation mode).
  • 0 papers support judge-vs-human agreement analysis.
  • 1 papers report explicit quality controls (calibration/adjudication/IAA).

Primary action: Start with the top 2 papers in “Start Here”, then validate assumptions in the protocol matrix.

Currently showing only replication-ready papers in ranking and matrix sections (6 papers).

Why This Matters For Eval Research

  • 21.1% of papers report explicit human-feedback signals, led by red-team protocols.
  • automatic metrics appears in 31.6% of papers in this hub.
  • AdvBench is a recurring benchmark anchor for cross-paper comparisons in this page.

Protocol Takeaways

  • Most common quality-control signal is rater calibration (1.8% of papers).
  • Rater context is mostly domain experts, and annotation is commonly trajectory-level annotation; use this to scope replication staffing.
  • Pair this hub with a human_eval-heavy hub to validate judge-model calibration.

Benchmark Interpretation

  • AdvBench appears in 3.5% of hub papers (2/57); use this cohort for benchmark-matched comparisons.
  • APPS appears in 1.8% of hub papers (1/57); use this cohort for benchmark-matched comparisons.

Metric Interpretation

  • jailbreak success rate is reported in 7% of hub papers (4/57); compare with a secondary metric before ranking methods.
  • accuracy is reported in 5.3% of hub papers (3/57); compare with a secondary metric before ranking methods.
Researcher Checklist (Expanded)

Researcher Checklist

  • Gap: Papers with explicit human feedback

    Coverage is a replication risk (21.1% vs 45% target).

  • Gap: Papers reporting quality controls

    Coverage is a replication risk (1.8% vs 30% target).

  • Gap: Papers naming benchmarks/datasets

    Coverage is a replication risk (5.3% vs 35% target).

  • Moderate: Papers naming evaluation metrics

    Coverage is usable but incomplete (24.6% vs 35% target).

  • Gap: Papers with known rater population

    Coverage is a replication risk (3.5% vs 35% target).

  • Gap: Papers with known annotation unit

    Coverage is a replication risk (5.3% vs 35% target).

Strengths

  • This hub still surfaces a concentrated paper set for protocol triage and replication planning.

Known Gaps

  • Only 1.8% of papers report quality controls; prioritize calibration/adjudication evidence.
  • Rater population is under-specified (3.5% coverage).
  • Annotation unit is under-specified (5.3% coverage).

Suggested Next Analyses

  • Pair this hub with a human_eval-heavy hub to validate judge-model calibration.
  • Stratify by benchmark (AdvBench vs APPS) before comparing methods.
  • Track metric sensitivity by reporting both jailbreak success rate and accuracy.
  • Add inter-annotator agreement checks when reproducing these protocols.
Recommended Queries (Expanded)

Recommended Queries

LLM-as-Judge Protocols Benchmark Slice: AdvBench Metric Slice: jailbreak success rate Recent High-Signal Papers

Start with These 3

Use these when you need one protocol anchor, one benchmark anchor, and one recent comparison point before reading the wider hub.

Start Here (Best First 6)

Ranked for protocol completeness (human signal, benchmark + metric anchors, quality controls, and judge/human overlap).

Protocol Matrix (Top 12)

Use this to quickly compare protocol ingredients instead of scanning long prose.

Paper HF Signal Eval Modes Benchmarks Metrics QC
Jailbreak Foundry: From Papers to Runnable Attacks for Reproducible Benchmarking

Feb 27, 2026

 Yes
Red Team
Llm As Judge AdvBench , Jbf Eval Success rate , Jailbreak success rate Not Reported
A Simple and Efficient Jailbreak Method Exploiting LLMs' Helpfulness

Sep 17, 2025

 Yes
Red Team
Automatic Metrics AdvBench Helpfulness Not Reported
RLShield: Practical Multi-Agent RL for Financial Cyber Defense with Attack-Surface MDPs and Real-Time Response Orchestration

Feb 26, 2026

 No
Not Reported
Automatic Metrics APPS Cost Not Reported
Weight space Detection of Backdoors in LoRA Adapters

Feb 16, 2026

 No
Not Reported
Automatic Metrics GSM8K , HumanEval+ Accuracy Not Reported
How Well Can LLM Agents Simulate End-User Security and Privacy Attitudes and Behaviors?

Feb 6, 2026

 No
Not Reported
Simulation Env Sp Abcbench Coherence Not Reported
PII-Bench: Evaluating Query-Aware Privacy Protection Systems

Feb 25, 2025

 No
Not Reported
Automatic Metrics Pii Bench Relevance Not Reported

Protocol Diff (Top Papers)

Fast side-by-side comparison for the highest-ranked papers in this hub.

Signal Jailbreak Foundry: From Papers to Runnable Attacks… A Simple and Efficient Jailbreak Method Exploiting… RLShield: Practical Multi-Agent RL for Financial Cy…
Human Feedback Red TeamRed TeamNot reported
Evaluation Modes Llm As JudgeAutomatic MetricsAutomatic Metrics
Benchmarks AdvBench, Jbf EvalAdvBenchAPPS
Metrics Success rate, Jailbreak success rateHelpfulnessCost
Quality Controls Not reportedNot reportedNot reported
Rater Population UnknownUnknownUnknown
Annotation Unit UnknownUnknownUnknown
Suggested Reading Order (Extended)

This section is intentionally expanded only when needed; use “Start Here” above for a faster pass.

Suggested Reading Order

  1. From Threat Intelligence to Firewall Rules: Semantic Relations in Hybrid AI Agent and Expert System Architectures

    Start here for detailed protocol reporting and quality-control evidence. Abstract: Web security demands rapid response capabilities to evolving cyber threats.

  2. On the Suitability of LLM-Driven Agents for Dark Pattern Audits

    Start here for detailed protocol reporting and quality-control evidence. Abstract: As LLM-driven agents begin to autonomously navigate the web, their ability to interpret and respond to manipulative interface.

  3. Contextualized Privacy Defense for LLM Agents

    Start here for detailed protocol reporting and quality-control evidence. Signals: simulation environments. Focus: helpfulness. Abstract: LLM agents increasingly act on users' personal information, yet existing privacy defenses remain.

  4. Jailbreak Foundry: From Papers to Runnable Attacks for Reproducible Benchmarking

    Include a human-eval paper to calibrate against judge-based evaluation settings. Signals: LLM-as-judge + red-team protocols. Focus: AdvBench / success rate. Abstract: This system enables a standardized AdvBench evaluation.

  5. What Matters For Safety Alignment?

    Include a human-eval paper to calibrate against judge-based evaluation settings. Signals: automatic metrics + red-team protocols. Focus: success rate. Abstract: This paper presents a comprehensive empirical study on.

  6. A Simple and Efficient Jailbreak Method Exploiting LLMs' Helpfulness

    Adds automatic metrics with red-team protocols for broader protocol coverage within this hub. Signals: automatic metrics + red-team protocols. Focus: AdvBench / helpfulness. Abstract: This study reveals a.

  7. Obscure but Effective: Classical Chinese Jailbreak Prompt Optimization via Bio-Inspired Search

    Adds automatic metrics with red-team protocols for broader protocol coverage within this hub. Signals: automatic metrics + red-team protocols. Focus: accuracy. Abstract: As Large Language Models (LLMs) are.

  8. MANATEE: Inference-Time Lightweight Diffusion Based Safety Defense for LLMs

    Adds automatic metrics with red-team protocols for broader protocol coverage within this hub. Signals: automatic metrics + red-team protocols. Focus: success rate. Abstract: Defending LLMs against adversarial jailbreak.

Known Limitations

Known Limitations

  • Only 1.8% of papers report quality controls; prioritize calibration/adjudication evidence.
  • Rater population is under-specified (3.5% coverage).
  • Narrative synthesis is grounded in metadata and abstracts only; full-paper implementation details are not parsed.
Research Utility Snapshot (Detailed)

Research Utility Snapshot

Human Feedback Mix

  • Red Team (11)
  • Expert Verification (1)

Evaluation Modes

  • Automatic Metrics (18)
  • Llm As Judge (2)
  • Simulation Env (2)

Top Benchmarks

  • AdvBench (2)
  • APPS (1)
  • Jbf Eval (1)

Top Metrics

  • Jailbreak success rate (4)
  • Accuracy (3)
  • Helpfulness (3)
  • Success rate (3)

Rater Population Mix

  • Domain Experts (2)

Quality Controls

  • Calibration (1)
Coverage diagnostics (sample-based): human-feedback 21.1% · benchmarks 12.3% · metrics 42.1% · quality controls 1.8%.

Top Papers

Related Hubs

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

Post a Job Get a Quote