Skip to content
← Back to explorer

HFEPX Hub

Long Horizon + Demonstrations Papers

Updated from current HFEPX corpus (Apr 9, 2026). 10 papers are grouped in this hub page.

Read Full Context

Updated from current HFEPX corpus (Apr 9, 2026). 10 papers are grouped in this hub page. Common evaluation modes: Simulation Env, Automatic Metrics. Most common rater population: Domain Experts. Common annotation unit: Trajectory. Frequently cited benchmark: Windowsagentarena. Common metric signal: cost. 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: 10 Last published: Mar 22, 2026 Global RSS Tag RSS
Long HorizonDemonstrations

Researcher Quick Triage

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

All Sampled Papers (10) Replication-Ready Only (1)

High-Signal Coverage

100.0%

10 / 10 sampled papers are not low-signal flagged.

Replication-Ready Set

1

Benchmark + metric + eval mode explicitly present.

Judge/Human Comparability

1

Papers containing both `human_eval` and `llm_as_judge`.

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

Primary action: Use this page for scouting only; collect additional papers before attempting replication-critical comparisons.

Currently showing only replication-ready papers in ranking and matrix sections (1 papers).

Need evaluators for this research workflow?

Post a Job →

Why This Matters For Eval Research

  • 100% of papers report explicit human-feedback signals, led by demonstration data.
  • simulation environments appears in 40% of papers in this hub.
  • Windowsagentarena is a recurring benchmark anchor for cross-paper comparisons in this page.

Protocol Takeaways

  • 1 sampled papers report both human evaluation and LLM-as-judge, supporting direct agreement checks.
  • Quality-control reporting is sparse in this slice; prioritize papers with explicit calibration or adjudication steps.
  • Rater context is mostly domain experts, and annotation is commonly trajectory-level annotation; use this to scope replication staffing.

Benchmark Interpretation

  • Windowsagentarena appears in 20% of hub papers (2/10); use this cohort for benchmark-matched comparisons.
  • OSWorld appears in 10% of hub papers (1/10); use this cohort for benchmark-matched comparisons.

Metric Interpretation

  • cost is reported in 10% of hub papers (1/10); compare with a secondary metric before ranking methods.
  • pass@1 is reported in 10% of hub papers (1/10); compare with a secondary metric before ranking methods.
Researcher Checklist (Expanded)

Researcher Checklist

  • Strong: Papers with explicit human feedback

    Coverage is strong (100% vs 45% target).

  • Gap: Papers reporting quality controls

    Coverage is a replication risk (0% vs 30% target).

  • Moderate: Papers naming benchmarks/datasets

    Coverage is usable but incomplete (30% vs 35% target).

  • Gap: Papers naming evaluation metrics

    Coverage is a replication risk (20% vs 35% target).

  • Gap: Papers with known rater population

    Coverage is a replication risk (20% vs 35% target).

  • Strong: Papers with known annotation unit

    Coverage is strong (80% vs 35% target).

Strengths

  • Strong human-feedback signal (100% of papers).
  • Contains both human-eval and LLM-as-judge protocols for head-to-head methodology comparison.
  • Agentic evaluation appears in 100% of papers.

Known Gaps

  • Only 0% of papers report quality controls; prioritize calibration/adjudication evidence.
  • Rater population is under-specified (20% coverage).
  • LLM-as-judge appears without enough inter-annotator agreement reporting.

Suggested Next Analyses

  • Compare papers that report both human_eval and llm_as_judge to quantify judge-human agreement drift.
  • Stratify by benchmark (Windowsagentarena vs OSWorld) before comparing methods.
  • Track metric sensitivity by reporting both cost and pass@1.
Recommended Queries (Expanded)

Recommended Queries

Judge vs Human Agreement Benchmark Slice: Windowsagentarena Metric Slice: cost 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
Suggested Reading Order (Extended)

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

Suggested Reading Order

  1. SOLE-R1: Video-Language Reasoning as the Sole Reward for On-Robot Reinforcement Learning

    Start here for detailed protocol reporting and quality-control evidence. Signals: simulation environments + demonstration data. Abstract: Vision-language models (VLMs) have shown impressive capabilities across diverse tasks, motivating efforts.

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

    Start here for detailed protocol reporting and quality-control evidence. Signals: human evaluation + demonstration data. Focus: WebArena / precision. Abstract: AgentHER realises this idea through a four-stage pipeline.

  3. RoboPocket: Improve Robot Policies Instantly with Your Phone

    Start here for detailed protocol reporting and quality-control evidence. Signals: demonstration data. Abstract: Scaling imitation learning is fundamentally constrained by the efficiency of data collection.

  4. RAPTOR: A Foundation Policy for Quadrotor Control

    Adds simulation environments with demonstration data for broader protocol coverage within this hub. Signals: simulation environments + demonstration data. Abstract: Humans are remarkably data-efficient when adapting to new.

  5. MoMaGen: Generating Demonstrations under Soft and Hard Constraints for Multi-Step Bimanual Mobile Manipulation

    Adds simulation environments with demonstration data for broader protocol coverage within this hub. Signals: simulation environments + demonstration data. Abstract: Imitation learning from large-scale, diverse human demonstrations has.

  6. Mastering Multi-Drone Volleyball through Hierarchical Co-Self-Play Reinforcement Learning

    Adds automatic metrics with demonstration data for broader protocol coverage within this hub. Signals: automatic metrics + demonstration data. Focus: win rate. Abstract: In this paper, we tackle.

  7. Watch and Learn: Learning to Use Computers from Online Videos

    Adds evaluation protocol evidence with demonstration data for broader protocol coverage within this hub. Signals: demonstration data. Focus: OSWorld. Abstract: Instead of directly generating actions or relying on.

  8. Efficient Agent Training for Computer Use

    Adds evaluation protocol evidence with demonstration data for broader protocol coverage within this hub. Signals: demonstration data. Focus: Windowsagentarena. Abstract: Scaling up high-quality trajectory data has long been.

Known Limitations

Known Limitations

  • Only 0% of papers report quality controls; prioritize calibration/adjudication evidence.
  • Rater population is under-specified (20% 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

  • Demonstrations (10)

Evaluation Modes

  • Simulation Env (4)
  • Automatic Metrics (1)
  • Human Eval (1)
  • Llm As Judge (1)

Top Benchmarks

  • Windowsagentarena (2)
  • OSWorld (1)
  • ToolBench (1)
  • WebArena (1)

Top Metrics

  • Cost (1)
  • Pass@1 (1)
  • Precision (1)
  • Win rate (1)

Rater Population Mix

  • Domain Experts (2)

Quality Controls

Coverage diagnostics (sample-based): human-feedback 100.0% · benchmarks 30.0% · metrics 20.0% · quality controls 0.0%.

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…

Related Hubs

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

Post a Job Get a Quote