Daily Archive
Updated from current HFEPX corpus (Feb 27, 2026). 6 papers are grouped in this daily page. Common evaluation modes: Automatic Metrics, Human Eval. Most common rater population: Domain Experts. Common annotation unit: Pairwise. Frequent quality control: Adjudication. Frequently cited benchmark: MMBench. Common metric signal: accuracy. Newest paper in this set is from Feb 15, 2026.
Updated from current HFEPX corpus (Feb 27, 2026). This page covers 6 papers centered on HFEPX Daily Archive: 2026-02-15. Common evaluation modes include Automatic Metrics, Human Eval, with benchmark emphasis on MMBench, Retrieval. Use the anchored takeaways below to compare protocol choices and identify papers with stronger evidence depth.
Evaluation emphasis: Automatic Metrics and Human Eval appear frequently in this slice.
Evidence: Does Socialization Emerge in AI Agent Society? A Case Study of Moltbook , MCPShield: A Security Cognition Layer for Adaptive Trust Calibration in Model Context Protocol Agents , Investigation for Relative Voice Impression Estimation
Benchmark concentration: MMBench, Retrieval helps control cross-paper variance.
Evidence: MCPShield: A Security Cognition Layer for Adaptive Trust Calibration in Model Context Protocol Agents , Investigation for Relative Voice Impression Estimation , Annotation-Efficient Vision-Language Model Adaptation to the Polish Language Using the LLaVA Framework
Metric concentration: accuracy, calibration is repeatedly reported in this group.
Evidence: Investigation for Relative Voice Impression Estimation , Annotation-Efficient Vision-Language Model Adaptation to the Polish Language Using the LLaVA Framework , Context Shapes LLMs Retrieval-Augmented Fact-Checking Effectiveness
Pair this hub with llm_as_judge pages to benchmark automated-vs-human evaluation tradeoffs.
Evidence: Annotation-Efficient Vision-Language Model Adaptation to the Polish Language Using the LLaVA Framework , Context Shapes LLMs Retrieval-Augmented Fact-Checking Effectiveness , HLE-Verified: A Systematic Verification and Structured Revision of Humanity's Last Exam
Stratify by benchmark (MMBench vs Retrieval) before comparing methods.
Evidence: Context Shapes LLMs Retrieval-Augmented Fact-Checking Effectiveness , HLE-Verified: A Systematic Verification and Structured Revision of Humanity's Last Exam , Does Socialization Emerge in AI Agent Society? A Case Study of Moltbook
Papers with explicit human feedback is visible in approximately 33.3% of papers in this set.
Evidence: HLE-Verified: A Systematic Verification and Structured Revision of Humanity's Last Exam , Does Socialization Emerge in AI Agent Society? A Case Study of Moltbook , MCPShield: A Security Cognition Layer for Adaptive Trust Calibration in Model Context Protocol Agents
Papers with explicit human feedback
Coverage is usable but incomplete (33.3% vs 45% target).
Papers reporting quality controls
Coverage is strong (33.3% vs 30% target).
Papers naming benchmarks/datasets
Coverage is usable but incomplete (33.3% vs 35% target).
Papers naming evaluation metrics
Coverage is strong (50% vs 35% target).
Papers with known rater population
Coverage is a replication risk (16.7% vs 35% target).
Papers with known annotation unit
Coverage is a replication risk (16.7% vs 35% target).
1. Does Socialization Emerge in AI Agent Society? A Case Study of Moltbook
Start with this anchor paper for scope and protocol framing. Covers Simulation Env.
Covers Automatic Metrics.
3. Investigation for Relative Voice Impression Estimation
Covers Automatic Metrics. Includes human-feedback signal: Pairwise Preference.
Covers Human Eval.
5. Context Shapes LLMs Retrieval-Augmented Fact-Checking Effectiveness
Covers Automatic Metrics.
6. HLE-Verified: A Systematic Verification and Structured Revision of Humanity's Last Exam
Covers Automatic Metrics. Includes human-feedback signal: Expert Verification, Critique Edit.
human_eval vs automatic_metrics
both=0, left_only=1, right_only=4
0 papers use both Human Eval and Automatic Metrics.
automatic_metrics vs simulation_env
both=0, left_only=4, 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.
Ming Li, Xirui Li, Tianyi Zhou · Feb 15, 2026
As large language model agents increasingly populate networked environments, a fundamental question arises: do artificial intelligence (AI) agent societies undergo convergence dynamics similar to human social systems?
Zhenhong Zhou, Yuanhe Zhang, Hongwei Cai, Moayad Aloqaily, Ouns Bouachir · Feb 15, 2026
The Model Context Protocol (MCP) standardizes tool use for LLM-based agents and enable third-party servers.
Kenichi Fujita, Yusuke Ijima · Feb 15, 2026
The estimation target is a low-dimensional vector derived from subjective evaluations, quantifying the perceptual shift of the second utterance relative to the first along an antonymic axis (e.g., ``Dark--Bright'').
Grzegorz Statkiewicz, Alicja Dobrzeniecka, Karolina Seweryn, Aleksandra Krasnodębska, Karolina Piosek · Feb 15, 2026
Despite relying almost entirely on automatic translation and minimal manual intervention to the training data, our approach yields strong results: we observe a +9.5% improvement over LLaVA-1.6-Vicuna-13B on a Polish-adapted MMBench, along w
Pietro Bernardelle, Stefano Civelli, Kevin Roitero, Gianluca Demartini · Feb 15, 2026
Large language models (LLMs) show strong reasoning abilities across diverse tasks, yet their performance on extended contexts remains inconsistent.
Weiqi Zhai, Zhihai Wang, Jinghang Wang, Boyu Yang, Xiaogang Li · Feb 15, 2026
Humanity's Last Exam (HLE) has become a widely used benchmark for evaluating frontier large language models on challenging, multi-domain questions.
