The paper presents EMPATH, a new multilingual multi-turn benchmark for safety evaluation of emotional-support chatbots that uses separate auditor and judge models and releases its pipeline and rubrics.
British Journal of Mathematical and Statistical Psychology61(1), 29–48 (2008)
10 Pith papers cite this work. Polarity classification is still indexing.
10
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
years
2026 10roles
background 2polarities
background 2representative citing papers
Fanfiction subgenres from AO3 function as universal register-based jailbreaks, raising mean attack success rate from 0.278 to 0.731 across eight aligned LLMs on HarmBench and JailbreakBench.
HTEB introduces dynamic, multi-axis evaluation of text embedding robustness using LLM transformations, finding decoupled profiles across models and that scaling does not close all robustness gaps.
RECOM dataset shows automatic metrics for open-ended Reddit QA exhibit a validity-discrimination tradeoff, with cosine similarity strong on validity but weak on model ranking, and BERTScore showing the reverse pattern after length control.
Introduces DFHR task, DFHR-Bench with over 180K triplets, and MFHC framework for mixed-modality dual face-hair retrieval.
A prompting method that forces GPAI models to state SE best practices before deciding reduces prompt-induced cognitive biases by 51% on average across eight tested biases.
Multimodal LLMs applied to 16 real-world configurators using 18 synthesized criteria can identify usability issues and generate actionable suggestions, with human review confirming reliability.
Analysis of 1,223 AI-HCI papers shows declining focus on human epistemic sovereignty and rising optimization of autonomous agents, leading to a proposal for scaffolded cognitive friction via multi-agent systems to preserve human cognitive agency.
citing papers explorer
-
EMPATH: A Multilingual Auditor-Judge Benchmark for Safety Evaluation of Emotional-Support Chatbots
The paper presents EMPATH, a new multilingual multi-turn benchmark for safety evaluation of emotional-support chatbots that uses separate auditor and judge models and releases its pipeline and rubrics.
-
Off-Distribution Voices: Fanfiction Subgenres as Universal Vernacular Jailbreaks for Aligned LLMs
Fanfiction subgenres from AO3 function as universal register-based jailbreaks, raising mean attack success rate from 0.278 to 0.731 across eight aligned LLMs on HarmBench and JailbreakBench.
-
The Harder Text Embedding Benchmark (HTEB): Beyond One-dimensional Static Robustness
HTEB introduces dynamic, multi-axis evaluation of text embedding robustness using LLM transformations, finding decoupled profiles across models and that scaling does not close all robustness gaps.
-
RECOM: A Validity Discrimination Tradeoff in Automatic Metrics for Open Ended Reddit Question Answering
RECOM dataset shows automatic metrics for open-ended Reddit QA exhibit a validity-discrimination tradeoff, with cosine similarity strong on validity but weak on model ranking, and BERTScore showing the reverse pattern after length control.
-
Mixed-Modality Dual Face-Hair Retrieval
Introduces DFHR task, DFHR-Bench with over 180K triplets, and MFHC framework for mixed-modality dual face-hair retrieval.
-
Mitigating Prompt-Induced Cognitive Biases in General-Purpose AI for Software Engineering
A prompting method that forces GPAI models to state SE best practices before deciding reduces prompt-induced cognitive biases by 51% on average across eight tested biases.
-
Usability Analysis of Configurator User Interfaces with Multimodal Large Language Models
Multimodal LLMs applied to 16 real-world configurators using 18 synthesized criteria can identify usability issues and generate actionable suggestions, with human review confirming reliability.
-
Cognitive Agency Surrender: Defending Epistemic Sovereignty via Scaffolded AI Friction
Analysis of 1,223 AI-HCI papers shows declining focus on human epistemic sovereignty and rising optimization of autonomous agents, leading to a proposal for scaffolded cognitive friction via multi-agent systems to preserve human cognitive agency.
- DialToM: A Theory of Mind Benchmark for Forecasting State-Driven Dialogue Trajectories
- Multilingual Training and Evaluation Resources for Vision-Language Models