# GNM Source: [GNM]() Collection index: [llms.txt]() A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. ## Benchmarks Page 1 of 2. This is an index page, not the full collection. - [Next index page](): Follow this link to discover more pages. - [Personalization on Community Alignment (CA)](): Personalization; Community Alignment (CA). - [Multi-perspective Alignment on Multi-Perspective Alignment Environment (test)](): Multi-perspective Alignment; Multi-Perspective Alignment Environment (test). - [Helpfulness alignment on HHH alignment](): Helpfulness alignment; HHH alignment. - [Quadratic Assignment Problem on QAP Cross-domain Transfer (nug12-30, tai20a-256c)](): Quadratic Assignment Problem; QAP (Quadratic Assignment Problem) Cross-domain Transfer nug12-30, tai20a-256c. - [Holistic Social Debiasing Assessment on Alignment and Bias Level Evaluation (ABLE)](): Holistic Social Debiasing Assessment; Alignment and Bias Level Evaluation (ABLE). - [Word Alignment on RWTH Gold Alignment de-en (test)](): Word Alignment; RWTH Gold Alignment de-en (test). - [Global-optimum attainment on Multi-drone assignment scenarios](): Global-optimum attainment; Multi-drone assignment scenarios. - [Community Alignment on Community Alignment (CA)](): Community Alignment; Community Alignment (CA). - [Honesty alignment on HHH alignment](): Honesty alignment; HHH alignment. - [Clinical Prediction on EHRSHOT Assignment of New Diag.](): Clinical Prediction; EHRSHOT Assignment of New Diag.. - [New Diagnosis Prediction on EHRSHOT Assignment of New Diagnoses](): New Diagnosis Prediction; EHRSHOT Assignment of New Diagnoses. - [Dual Face-Hair Retrieval on DFHR-Bench Bimodal Alignment](): Dual Face-Hair Retrieval; DFHR-Bench Bimodal Alignment. - [General Alignment on General Alignment and Coherence](): General Alignment; General Alignment and Coherence. - [Spatial Alignment on spatial alignment benchmark](): spatial alignment; spatial alignment benchmark. - [Deception Detection on Agentic Misalignment](): Deception Detection; Agentic Misalignment. - [Specification Alignment on Specification Alignment Evaluation Set (full (1500))](): Specification Alignment; Specification Alignment Evaluation Set (full (1500)). - [Model Alignment Evaluation on Stoic Alignment Benchmark](): Model Alignment Evaluation; Stoic Alignment Benchmark. - [ASR-OCR Alignment on Technical Presentation Slide Alignment Dataset 96 samples 1.0](): ASR-OCR Alignment; Technical Presentation Slide Alignment Dataset 96 samples 1.0. - [Generalist Alignment on AlpacaEval 2.0](): Generalist Alignment; AlpacaEval Generalist alignment prompts 2.0. - [Chinese Value Alignment Benchmark 1.0 (Chinese)](): Value Alignment Evaluation; Chinese Value Alignment Benchmark 1.0 (Chinese). - [Latent Knowledge Elicitation on Deceptive Alignment Benchmark (DAB) 400 scenarios](): Latent Knowledge Elicitation; Deceptive Alignment Benchmark (DAB) 400 scenarios. - [Prompt Hijacking on IHEval Prompt Hijacking - Alignment 1.0](): Prompt Hijacking; IHEval Prompt Hijacking - Alignment 1.0. - [Prompt Extraction on IHEval Prompt Extraction Alignment 1.0](): Prompt Extraction; IHEval Prompt Extraction Alignment 1.0. - [Safety Alignment on Safety Alignment](): Safety Alignment; Safety Alignment. - [Novelty Alignment unseen (test)](): Novelty Alignment; Novelty Alignment unseen (test). - [Music Textual Alignment on M-TA](): Music Textual Alignment; M-TA (Music Textual Alignment). - [Value Alignment Text Rewriting on Value Alignment Dataset](): Value Alignment Text Rewriting; Value Alignment Dataset. - [Text-to-Image Generation on Text-to-Image Alignment Prompts FLUX.2-Klein-4B (test)](): Text-to-Image Generation; Text-to-Image Alignment Prompts FLUX.2-Klein-4B (test). - [Lifelong Alignment on Six sequential tasks for lifelong alignment](): Lifelong Alignment; Six sequential tasks for lifelong alignment. - [Part-level Alignment on Part-level Alignment Base](): Part-level Alignment; Part-level Alignment Base. - [Part-level Alignment on Part-level Alignment (All Pick5)](): Part-level Alignment; Part-level Alignment (All Pick5). - [Part-level Alignment on Part-level Alignment (All)](): Part-level Alignment; Part-level Alignment (All). - [Cultural Awareness & Instruction Following on Arabic Cultural Value Alignment (test)](): Cultural Awareness & Instruction Following; Arabic Cultural Value Alignment (test). - [Safety Alignment on Safety Alignment Dataset 4-order (test)](): Safety Alignment; Safety Alignment Dataset 4-order (test). - [Safety Alignment on Safety Alignment Dataset 3-order (test)](): Safety Alignment; Safety Alignment Dataset 3-order (test). - [Safety Alignment on Safety Alignment Dataset 2-order (test)](): Safety Alignment; Safety Alignment Dataset 2-order (test). - [Safety Alignment on Safety Alignment Dataset 1-order (test)](): Safety Alignment; Safety Alignment Dataset 1-order (test). - [Dynamic Real2Sim alignment on Real2Sim alignment dataset (test)](): Dynamic Real2Sim alignment; Real2Sim alignment dataset (Milk, Oreo, Ovaltine, Spam) 1.0 (test). - [Link Prediction on GNM-30K (test)](): Link Prediction; GNM-30K (test). - [3D Object Detection on nuScenes Noise Misalignment (val)](): 3D Object Detection; nuScenes Noise Misalignment (val). - [Dialogue Generation on Dialogue-level evaluation Ethical and Affective Alignment](): Dialogue Generation; Dialogue-level evaluation Ethical and Affective Alignment. - [Reward Modeling on HHH-Alignment Reversed](): Reward Modeling; HHH-Alignment Reversed. - [Reward Modeling on HHH-Alignment Standard](): Reward Modeling; HHH-Alignment Standard. - [Clustering on Hdigit 1% alignment (test)](): Clustering; Hdigit 1% alignment (test). - [Clustering on MNIST-USPS 1% alignment (test)](): Clustering; MNIST-USPS 1% alignment (test). - [Moral Alignment on Moral Alignment](): Moral Alignment; Moral Alignment. - [Helpfulness-Safety Alignment on PKU-Alignment](): Helpfulness-Safety Alignment; PKU-Alignment. - [Text-to-Image Generation on General Prompts Alignment & Preference](): Text-to-Image Generation; General Prompts Alignment & Preference. - [Text-to-Image Generation on Stable Diffusion Alignment Prompts 1.5 (test)](): Text-to-Image Generation; Stable Diffusion Alignment Prompts 1.5 (test). - [Clustering on BDGP 1% alignment (test)](): Clustering; BDGP 1% alignment (test). - [Clustering on HandWritten 1% alignment (test)](): Clustering; HandWritten 1% alignment (test). - [Reward Modeling on HHH-Alignment (OOD)](): Reward Modeling; HHH-Alignment (OOD). - [Reward Modeling on HHH-Alignment OOD (test)](): Reward Modeling; HHH-Alignment OOD (test). - [Assignment Problem](): Assignment Problem; Assignment Problem. - [Clustering on yale\_mtv alignment rate 0.3, incomplete rate 0.5](): Clustering; yale\_mtv alignment rate 0.3, incomplete rate 0.5. - [Clustering on ALOI alignment rate 0.3, incomplete rate 0.5](): Clustering; ALOI alignment rate 0.3, incomplete rate 0.5. - [Clustering on MSRC alignment rate 0.3, incomplete rate 0.5 v1](): Clustering; MSRC alignment rate 0.3, incomplete rate 0.5 v1. - [Clustering on Reuters dim10 (alignment rate 0.3, incomplete rate 0.5)](): Clustering; Reuters dim10 (alignment rate 0.3, incomplete rate 0.5). - [Clustering on BDGP alignment rate 0.3, incomplete rate 0.5](): Clustering; BDGP alignment rate 0.3, incomplete rate 0.5. - [Clustering on 3Sources alignment rate 0.3, incomplete rate 0.5](): Clustering; 3Sources alignment rate 0.3, incomplete rate 0.5. - [Personalized Image Generation on Demographic Alignment Dataset](): Personalized Image Generation; Demographic Alignment Dataset. - [Value-behavior alignment on Schwartz Value-Behavior Alignment Charity, Donation, Prosocial, Everyday, Big Five](): Value-behavior alignment; Schwartz Value-Behavior Alignment Charity, Donation, Prosocial, Everyday, Big Five. - [Text-to-Image Alignment on alignment prompts SDXL](): Text-to-Image Alignment; alignment prompts SDXL. - [AIGC Video Quality Assessment on FETV Alignment aspect (test)](): AIGC Video Quality Assessment; FETV Alignment aspect (test). - [Human Preference Alignment Out-of-Domain (test)](): Human preference alignment; Human Preference Alignment Out-of-Domain (test). - [Human Preference Alignment on In-Domain (test)](): Human preference alignment; Human Preference Alignment In-Domain (test). - [Video Super-Resolution on Synthetic Video Alignment Benchmark (test)](): Video Super-Resolution; Synthetic Video Alignment Benchmark (test). - [Misalignment monitoring on Misalignment Monitoring Dataset (test)](): Misalignment monitoring; Misalignment Monitoring Dataset (test). - [Misalignment Monitoring Dataset (dev)](): Misalignment monitoring; Misalignment Monitoring Dataset (dev). - [Ethical Alignment on 24 questions emergent misalignment scenario (test)](): Ethical Alignment; 24 questions emergent misalignment scenario (test). - [Low-Frequency Alignment on Photomosaic Alignment Evaluation Dataset (test)](): Low-Frequency Alignment; Photomosaic Alignment Evaluation Dataset (test). - [Function Addition on MMLU Safety Alignment](): Function Addition; MMLU Safety Alignment. - [Word Alignment](): Word alignment; Word alignment. - [Cultural Alignment Assessment on Cultural alignment dataset](): Cultural alignment assessment; Cultural alignment dataset. - [Combinatorial Reinforcement Learning on Dynamic Assignment](): Combinatorial Reinforcement Learning; Dynamic Assignment. - [Text-to-Image Generation on General Image Quality and Alignment Evaluation](): Text-to-Image Generation; General Image Quality and Alignment Evaluation. - [Human Preference Alignment Evaluation](): Human Preference Alignment; Human Preference Alignment. - [Heading angle estimation on EvalR1-EvalR4 120 Seconds alignment (evaluation)](): Heading angle estimation; EvalR1-EvalR4 120 Seconds alignment (evaluation). - [Heading angle estimation on EvalR1-EvalR4 90 Seconds alignment (test)](): Heading angle estimation; EvalR1-EvalR4 90 Seconds alignment (test). - [Heading angle estimation on EvalR1-EvalR4 60 Seconds alignment (Evaluation)](): Heading angle estimation; EvalR1-EvalR4 60 Seconds alignment (Evaluation).