HFEPX Hub
Updated from current HFEPX corpus (Feb 27, 2026). 11 papers are grouped in this hub page. Common evaluation modes: Automatic Metrics, Human Eval. Common annotation unit: Scalar. Frequently cited benchmark: AIME. 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 Feb 26, 2026.
Updated from current HFEPX corpus (Feb 27, 2026). This page tracks 11 papers for CS.AI + Critique Edit Papers. Dominant protocol signals include automatic metrics, human evaluation, simulation environments, with frequent benchmark focus on AIME, Contentbench and metric focus on cost, accuracy. Use the grounded sections below to prioritize reproducible protocol choices, benchmark-matched comparisons, and judge-vs-human evaluation checks.
100% of papers report explicit human-feedback signals, led by critique/edit feedback.
Evidence: Natural Language Declarative Prompting (NLD-P): A Modular Governance Method for Prompt Design Under Model Drift , Beyond Refusal: Probing the Limits of Agentic Self-Correction for Semantic Sensitive Information , CAMEL: Confidence-Gated Reflection for Reward Modeling , Can Large Language Models Replace Human Coders? Introducing ContentBench
automatic metrics appears in 81.8% of papers in this hub.
Evidence: Natural Language Declarative Prompting (NLD-P): A Modular Governance Method for Prompt Design Under Model Drift , Beyond Refusal: Probing the Limits of Agentic Self-Correction for Semantic Sensitive Information , CAMEL: Confidence-Gated Reflection for Reward Modeling , Can Large Language Models Replace Human Coders? Introducing ContentBench
AIME is a recurring benchmark anchor for cross-paper comparisons in this page.
Evidence: Critique-GRPO: Advancing LLM Reasoning with Natural Language and Numerical Feedback , Natural Language Declarative Prompting (NLD-P): A Modular Governance Method for Prompt Design Under Model Drift , Beyond Refusal: Probing the Limits of Agentic Self-Correction for Semantic Sensitive Information , CAMEL: Confidence-Gated Reflection for Reward Modeling
Quality-control reporting is sparse in this slice; prioritize papers with explicit calibration or adjudication steps.
Evidence: Natural Language Declarative Prompting (NLD-P): A Modular Governance Method for Prompt Design Under Model Drift , Beyond Refusal: Probing the Limits of Agentic Self-Correction for Semantic Sensitive Information , CAMEL: Confidence-Gated Reflection for Reward Modeling , Can Large Language Models Replace Human Coders? Introducing ContentBench
Rater context is mostly unspecified rater pools, and annotation is commonly scalar scoring; use this to scope replication staffing.
Evidence: Natural Language Declarative Prompting (NLD-P): A Modular Governance Method for Prompt Design Under Model Drift , Beyond Refusal: Probing the Limits of Agentic Self-Correction for Semantic Sensitive Information , CAMEL: Confidence-Gated Reflection for Reward Modeling , Can Large Language Models Replace Human Coders? Introducing ContentBench
Pair this hub with llm_as_judge pages to benchmark automated-vs-human evaluation tradeoffs.
Evidence: RebuttalAgent: Strategic Persuasion in Academic Rebuttal via Theory of Mind , Natural Language Declarative Prompting (NLD-P): A Modular Governance Method for Prompt Design Under Model Drift , Beyond Refusal: Probing the Limits of Agentic Self-Correction for Semantic Sensitive Information , CAMEL: Confidence-Gated Reflection for Reward Modeling
Papers with explicit human feedback
Coverage is strong (100% vs 45% target).
Papers reporting quality controls
Coverage is a replication risk (0% vs 30% target).
Papers naming benchmarks/datasets
Coverage is strong (36.4% vs 35% target).
Papers naming evaluation metrics
Coverage is strong (36.4% vs 35% target).
Papers with known rater population
Coverage is a replication risk (0% vs 35% target).
Papers with known annotation unit
Coverage is a replication risk (9.1% vs 35% target).
Start here for detailed protocol reporting, including rater and quality-control evidence.
2. Beyond Refusal: Probing the Limits of Agentic Self-Correction for Semantic Sensitive Information
Start here for detailed protocol reporting, including rater and quality-control evidence.
3. CAMEL: Confidence-Gated Reflection for Reward Modeling
Start here for detailed protocol reporting, including rater and quality-control evidence.
4. RebuttalAgent: Strategic Persuasion in Academic Rebuttal via Theory of Mind
Include a human-eval paper to anchor calibration against automated judge settings.
5. Can Large Language Models Replace Human Coders? Introducing ContentBench
Adds automatic metrics with critique/edit feedback for broader coverage within this hub.
6. Unlocking Reasoning Capability on Machine Translation in Large Language Models
Adds automatic metrics with critique/edit feedback for broader coverage within this hub.
Adds simulation environments with critique/edit feedback for broader coverage within this hub.
8. SocialHarmBench: Revealing LLM Vulnerabilities to Socially Harmful Requests
Adds automatic metrics with critique/edit feedback for broader coverage within this hub.
human_eval vs automatic_metrics
both=0, left_only=1, right_only=9
0 papers use both Human Eval and Automatic Metrics.
automatic_metrics vs simulation_env
both=0, left_only=9, right_only=1
0 papers use both Automatic Metrics and Simulation Env.
human_eval vs simulation_env
both=0, left_only=1, right_only=1
0 papers use both Human Eval and Simulation Env.
Benchmark Brief
AIME
Coverage: 1 papers (9.1%)
1 papers (9.1%) mention AIME.
Examples: Critique-GRPO: Advancing LLM Reasoning with Natural Language and Numerical Feedback
Benchmark Brief
Contentbench
Coverage: 1 papers (9.1%)
1 papers (9.1%) mention Contentbench.
Examples: Can Large Language Models Replace Human Coders? Introducing ContentBench
Benchmark Brief
Rebuttalbench
Coverage: 1 papers (9.1%)
1 papers (9.1%) mention Rebuttalbench.
Examples: RebuttalAgent: Strategic Persuasion in Academic Rebuttal via Theory of Mind
Metric Brief
cost
Coverage: 2 papers (18.2%)
2 papers (18.2%) mention cost.
Examples: CAMEL: Confidence-Gated Reflection for Reward Modeling , Can Large Language Models Replace Human Coders? Introducing ContentBench
Metric Brief
accuracy
Coverage: 1 papers (9.1%)
1 papers (9.1%) mention accuracy.
Examples: CAMEL: Confidence-Gated Reflection for Reward Modeling
Metric Brief
agreement
Coverage: 1 papers (9.1%)
1 papers (9.1%) mention agreement.
Examples: Can Large Language Models Replace Human Coders? Introducing ContentBench
Most Cited In This Hub
Fast path to methods with the strongest citation traction in this scope.
Papers: Natural Language Declarative Prompting (NLD-P): A Modular Governance Method for Prompt Design Under Model Drift , Beyond Refusal: Probing the Limits of Agentic Self-Correction for Semantic Sensitive Information , CAMEL: Confidence-Gated Reflection for Reward Modeling
Most Recent
Fast path to latest protocol changes and newly published evaluation setups.
Papers: Natural Language Declarative Prompting (NLD-P): A Modular Governance Method for Prompt Design Under Model Drift , Beyond Refusal: Probing the Limits of Agentic Self-Correction for Semantic Sensitive Information , CAMEL: Confidence-Gated Reflection for Reward Modeling
Best Protocol Detail
Papers with explicit rater/unit metadata and quality-control signals for reproducibility.
Papers: Natural Language Declarative Prompting (NLD-P): A Modular Governance Method for Prompt Design Under Model Drift , Beyond Refusal: Probing the Limits of Agentic Self-Correction for Semantic Sensitive Information , CAMEL: Confidence-Gated Reflection for Reward Modeling
Hyunwoo Kim, Hanau Yi, Jaehee Bae, Yumin Kim · Feb 26, 2026 · Citations: 0
NLD-P is formalized as a modular control abstraction that separates provenance, constraint logic, task content, and post-generation evaluation, encoded directly in natural language without reliance on external orchestration code.
Umid Suleymanov, Zaur Rajabov, Emil Mirzazada, Murat Kantarcioglu · Feb 25, 2026 · Citations: 0
To address this, we introduce SemSIEdit, an inference-time framework where an agentic "Editor" iteratively critiques and rewrites sensitive spans to preserve narrative flow rather than simply refusing to answer.
Zirui Zhu, Hailun Xu, Yang Luo, Yong Liu, Kanchan Sarkar · Feb 24, 2026 · Citations: 0
Reward models play a fundamental role in aligning large language models with human preferences.
Michael Haman · Feb 23, 2026 · Citations: 0
This paper introduces ContentBench, a public benchmark suite that helps answer this replacement question by tracking how much agreement low-cost LLMs achieve and what they cost on the same interpretive coding tasks.
Sara Rajaee, Sebastian Vincent, Alexandre Berard, Marzieh Fadaee, Kelly Marchisio · Feb 16, 2026 · Citations: 0
We systematically evaluate several open- and closed-weights RLMs on the WMT24++ benchmark and find that enabling explicit reasoning consistently degrades translation quality across languages and models.
Sha Li, Stefano Petrangeli, Yu Shen, Xiang Chen · Feb 14, 2026 · Citations: 0
We introduce LaySPA, a reinforcement learning framework that equips large language models (LLMs) with explicit and interpretable spatial reasoning for content-aware graphic layout design.
Zhitao He, Zongwei Lyu, Yi R Fung · Jan 22, 2026 · Citations: 0
In this paper, we introduce RebuttalAgent, the first framework to ground academic rebuttal in Theory of Mind (ToM), operationalized through a ToM-Strategy-Response (TSR) framework that models reviewer mental state, formulates persuasion str
Punya Syon Pandey, Hai Son Le, Devansh Bhardwaj, Rada Mihalcea, Zhijing Jin · Oct 6, 2025 · Citations: 0
Yet, existing safety benchmarks rarely test vulnerabilities in domains such as political manipulation, propaganda and disinformation generation, or surveillance and information control.
Kihoon Son, DaEun Choi, Tae Soo Kim, Young-Ho Kim, Sangdoo Yun · Sep 18, 2025 · Citations: 0
Furthermore, exploratory applications demonstrate that captured steps can enhance generative AI agents in Figma, yielding predictions better aligned with professionals and producing coherent outcomes.
Nicole Cho, Kirsty Fielding, William Watson, Sumitra Ganesh, Manuela Veloso · Aug 18, 2025 · Citations: 0
To address this, we present TASER (Table Agents for Schema-guided Extraction and Recommendation), a continuously learning, agentic table extraction system that converts highly unstructured, multi-page, heterogeneous tables into normalized,
Xiaoying Zhang, Yipeng Zhang, Hao Sun, Kaituo Feng, Chaochao Lu · Jun 3, 2025 · Citations: 0
Recent advances in reinforcement learning (RL) using numerical rewards have significantly enhanced the complex reasoning capabilities of large language models (LLMs).
