Skip to content
← Back to explorer

HFEPX Hub

CS.CL + Human Eval Papers

Updated from current HFEPX corpus (Apr 12, 2026). 84 papers are grouped in this hub page.

Read Full Context

Updated from current HFEPX corpus (Apr 12, 2026). 84 papers are grouped in this hub page. Common evaluation modes: Human Eval, Automatic Metrics. Most common rater population: Domain Experts. Common annotation unit: Multi Dim Rubric. Frequent quality control: Inter Annotator Agreement Reported. Frequently cited benchmark: Rewardbench. Common metric signal: accuracy. 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 Mar 22, 2026.

Papers: 84 Last published: Mar 22, 2026 Global RSS Tag RSS
Cs.CLHuman Eval

Researcher Quick Triage

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

Analysis blocks below are computed from the currently loaded sample (60 of 84 total papers in this hub).

All Sampled Papers (60) Replication-Ready Only (6)

High-Signal Coverage

100.0%

60 / 60 sampled papers are not low-signal flagged.

Replication-Ready Set

6

Benchmark + metric + eval mode explicitly present.

Judge/Human Comparability

10

Papers containing both `human_eval` and `llm_as_judge`.

  • 6 papers are replication-ready (benchmark + metric + explicit evaluation mode).
  • 10 papers support judge-vs-human agreement analysis.
  • 7 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).

Need evaluators for this research workflow?

Post a Job →

Why This Matters For Eval Research

  • 41.8% of papers report explicit human-feedback signals, led by pairwise preferences.
  • human evaluation appears in 65.5% of papers in this hub.
  • Rewardbench is a recurring benchmark anchor for cross-paper comparisons in this page.

Protocol Takeaways

  • 3 sampled papers report both human evaluation and LLM-as-judge, supporting direct agreement checks.
  • Most common quality-control signal is inter-annotator agreement reporting (4.8% of papers).
  • Rater context is mostly domain experts, and annotation is commonly multi-dimensional rubrics; use this to scope replication staffing.

Benchmark Interpretation

  • Rewardbench appears in 3.6% of hub papers (2/84); use this cohort for benchmark-matched comparisons.
  • AIME appears in 1.8% of hub papers (1/84); use this cohort for benchmark-matched comparisons.

Metric Interpretation

  • accuracy is reported in 27.3% of hub papers (15/84); compare with a secondary metric before ranking methods.
  • agreement is reported in 14.5% of hub papers (8/84); compare with a secondary metric before ranking methods.
Researcher Checklist (Expanded)

Researcher Checklist

  • Moderate: Papers with explicit human feedback

    Coverage is usable but incomplete (41.8% vs 45% target).

  • Gap: Papers reporting quality controls

    Coverage is a replication risk (12.7% vs 30% target).

  • Moderate: Papers naming benchmarks/datasets

    Coverage is usable but incomplete (27.3% vs 35% target).

  • Strong: Papers naming evaluation metrics

    Coverage is strong (67.3% vs 35% target).

  • Moderate: Papers with known rater population

    Coverage is usable but incomplete (30.9% vs 35% target).

  • Moderate: Papers with known annotation unit

    Coverage is usable but incomplete (27.3% vs 35% target).

Strengths

  • Contains both human-eval and LLM-as-judge protocols for head-to-head methodology comparison.

Known Gaps

  • Only 12.7% of papers report quality controls; prioritize calibration/adjudication evidence.

Suggested Next Analyses

  • Compare papers that report both human_eval and llm_as_judge to quantify judge-human agreement drift.
  • Stratify by benchmark (Rewardbench vs AIME) before comparing methods.
  • Track metric sensitivity by reporting both accuracy and agreement.
Recommended Queries (Expanded)

Recommended Queries

Judge vs Human Agreement Benchmark Slice: Rewardbench Metric Slice: accuracy IAA-Reported Evaluations 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
AgentHER: Hindsight Experience Replay for LLM Agent Trajectory Relabeling

Mar 22, 2026

 Yes
Demonstrations
Human Eval , Llm As Judge WebArena , ToolBench Precision , Pass@1 Not Reported
PoSh: Using Scene Graphs To Guide LLMs-as-a-Judge For Detailed Image Descriptions

Oct 21, 2025

 Yes
Rubric Rating
Human Eval , Llm As Judge CAPArena Spearman Not Reported
Personalized RewardBench: Evaluating Reward Models with Human Aligned Personalization

Apr 8, 2026

 Yes
Pairwise Preference , Rubric Rating
Human Eval , Automatic Metrics Rewardbench Accuracy , Helpfulness Not Reported
LMUnit: Fine-grained Evaluation with Natural Language Unit Tests

Dec 17, 2024

 Yes
Pairwise Preference
Human Eval Biggenbench , Rewardbench Agreement Inter Annotator Agreement Reported
DataSTORM: Deep Research on Large-Scale Databases using Exploratory Data Analysis and Data Storytelling

Apr 7, 2026

 No
Not Reported
Human Eval Insightbench Recall Not Reported
Beyond Rows to Reasoning: Agentic Retrieval for Multimodal Spreadsheet Understanding and Editing

Mar 6, 2026

 No
Not Reported
Human Eval , Automatic Metrics Frtr Bench Accuracy , Cost Not Reported

Protocol Diff (Top Papers)

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

Signal AgentHER: Hindsight Experience Replay for LLM Agent… PoSh: Using Scene Graphs To Guide LLMs-as-a-Judge F… Personalized RewardBench: Evaluating Reward Models…
Human Feedback DemonstrationsRubric RatingPairwise Preference, Rubric Rating
Evaluation Modes Human Eval, Llm As JudgeHuman Eval, Llm As JudgeHuman Eval, Automatic Metrics
Benchmarks WebArena, ToolBenchCAPArenaRewardbench
Metrics Precision, Pass@1SpearmanAccuracy, Helpfulness
Quality Controls Not reportedNot reportedNot reported
Rater Population UnknownDomain ExpertsUnknown
Annotation Unit TrajectoryMulti Dim RubricPairwise
Suggested Reading Order (Extended)

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

Suggested Reading Order

  1. CounselReflect: A Toolkit for Auditing Mental-Health Dialogues

    Start here for detailed protocol reporting and quality-control evidence. Signals: human evaluation + rubric ratings. Abstract: The system integrates two families of evaluation signals: (i) 12 model-based metrics.

  2. Is this Idea Novel? An Automated Benchmark for Judgment of Research Ideas

    Start here for detailed protocol reporting and quality-control evidence. Signals: human evaluation + rubric ratings. Focus: Rinobench. Abstract: Yet, evaluation of these approaches remains largely inconsistent and is.

  3. Personalized RewardBench: Evaluating Reward Models with Human Aligned Personalization

    High citation traction makes this a strong baseline for protocol comparison. Signals: human evaluation + pairwise preferences. Focus: Rewardbench / accuracy. Abstract: While benchmarks for general response quality.

  4. Sell More, Play Less: Benchmarking LLM Realistic Selling Skill

    High citation traction makes this a strong baseline for protocol comparison. Signals: human evaluation. Abstract: We propose a fully automatic evaluation pipeline that combines (i) an LLM-based rater.

  5. How Long Reasoning Chains Influence LLMs' Judgment of Answer Factuality

    High citation traction makes this a strong baseline for protocol comparison. Signals: human evaluation. Focus: accuracy. Abstract: Large language models (LLMs) has been widely adopted as a scalable.

  6. AgentHER: Hindsight Experience Replay for LLM Agent Trajectory Relabeling

    Include a human-eval paper to calibrate against judge-based evaluation settings. Signals: human evaluation + demonstration data. Focus: WebArena / precision. Abstract: AgentHER realises this idea through a four-stage.

  7. PoSh: Using Scene Graphs To Guide LLMs-as-a-Judge For Detailed Image Descriptions

    Include an LLM-as-judge paper to test judge design and agreement assumptions. Signals: human evaluation + rubric ratings. Focus: CAPArena / spearman. Abstract: In this work, we introduce PoSh,.

  8. LMUnit: Fine-grained Evaluation with Natural Language Unit Tests

    Adds human evaluation with pairwise preferences for broader protocol coverage within this hub. Signals: human evaluation + pairwise preferences. Focus: Biggenbench / agreement. Abstract: As language models become.

Known Limitations

Known Limitations

  • Only 12.7% of papers report quality controls; prioritize calibration/adjudication evidence.
  • Narrative synthesis is grounded in metadata and abstracts only; full-paper implementation details are not parsed.
  • Cross-page comparisons should be benchmark- and metric-matched to avoid protocol confounding.
Research Utility Snapshot (Detailed)

Research Utility Snapshot

Human Feedback Mix

  • Pairwise Preference (12)
  • Rubric Rating (9)
  • Expert Verification (4)
  • Critique Edit (3)

Evaluation Modes

  • Human Eval (55)
  • Automatic Metrics (21)
  • Llm As Judge (10)
  • Simulation Env (4)

Top Benchmarks

  • Rewardbench (2)
  • AIME (1)
  • APPS (1)
  • Biggenbench (1)

Top Metrics

  • Accuracy (15)
  • Agreement (8)
  • F1 (5)
  • Bleu (4)

Rater Population Mix

  • Domain Experts (15)
  • Mixed (2)

Quality Controls

  • Inter Annotator Agreement Reported (4)
  • Adjudication (3)
  • Gold Questions (2)
Coverage diagnostics (sample-based): human-feedback 38.3% · benchmarks 25.0% · metrics 61.7% · quality controls 11.7%.

Top Papers

  • AgentHER: Hindsight Experience Replay for LLM Agent Trajectory Relabeling

    Liang Ding · Mar 22, 2026 · Citations: 0

    Demonstrations Human EvalLlm As Judge Long Horizon

    LLM agents fail on the majority of real-world tasks -- GPT-4o succeeds on fewer than 15% of WebArena navigation tasks and below 55% pass@1 on ToolBench (Zhou et al., 2024; Qin et al., 2024) -- yet every failed trajectory is routinely…

  • LMUnit: Fine-grained Evaluation with Natural Language Unit Tests

    Jon Saad-Falcon, Rajan Vivek, William Berrios, Nandita Shankar Naik, Matija Franklin · Dec 17, 2024 · Citations: 0

    Pairwise Preference Human Eval

    We introduce natural language unit tests, a paradigm that decomposes response quality into explicit, testable criteria, along with a unified scoring model, LMUnit, which combines multi-objective training across preferences, direct ratings,…

  • PoSh: Using Scene Graphs To Guide LLMs-as-a-Judge For Detailed Image Descriptions

    Amith Ananthram, Elias Stengel-Eskin, Lorena A. Bradford, Julia Demarest, Adam Purvis · Oct 21, 2025 · Citations: 0

    Rubric Rating Human EvalLlm As Judge

    In this work, we introduce PoSh, a metric for detailed image description that uses scene graphs as structured rubrics to guide LLMs-as-a-Judge, producing aggregate scores grounded in fine-grained errors (e.g.

  • Personalized RewardBench: Evaluating Reward Models with Human Aligned Personalization

    Qiyao Ma, Dechen Gao, Rui Cai, Boqi Zhao, Hanchu Zhou · Apr 8, 2026 · Citations: 0

    Pairwise PreferenceRubric Rating Human EvalAutomatic Metrics

    Pluralistic alignment has emerged as a critical frontier in the development of Large Language Models (LLMs), with reward models (RMs) serving as a central mechanism for capturing diverse human values.

  • Beyond Rows to Reasoning: Agentic Retrieval for Multimodal Spreadsheet Understanding and Editing

    Anmol Gulati, Sahil Sen, Waqar Sarguroh, Kevin Paul · Mar 6, 2026 · Citations: 0

    Human EvalAutomatic Metrics Long Horizon

    We introduce Beyond Rows to Reasoning (BRTR), a multimodal agentic framework for spreadsheet understanding that replaces single-pass retrieval with an iterative tool-calling loop, supporting end-to-end Excel workflows from complex analysis…

  • DataSTORM: Deep Research on Large-Scale Databases using Exploratory Data Analysis and Data Storytelling

    Shicheng Liu, Yucheng Jiang, Sajid Farook, Camila Nicollier Sanchez, David Fernando Castro Pena · Apr 7, 2026 · Citations: 0

    Human Eval Long Horizon

    Deep research with Large Language Model (LLM) agents is emerging as a powerful paradigm for multi-step information discovery, synthesis, and analysis.

Related Hubs

Get Started

Join the #1 Platform for AI Training Talent

Where top AI builders and expert AI Trainers connect to build the future of AI.
Self-Service
Post a Job
Post your project and get a shortlist of qualified AI Trainers and Data Labelers. Hire and manage your team in the tools you already use.
Create Account & Post a Job
Managed Service
For Large Projects
Done-for-You
We recruit, onboard, and manage a dedicated team inside your tools. End-to-end operations for large or complex projects.
Learn About Managed Service
For Freelancers
Join as an AI Trainer
Find AI training and data labeling projects across platforms, all in one place. One profile, one application process, more opportunities.
Join Now