https://evalsuite-nine.vercel.app https://evalsuite-nine.vercel.app/playground https://evalsuite-nine.vercel.app/benchmarks https://evalsuite-nine.vercel.app/research https://evalsuite-nine.vercel.app/roadmap https://evalsuite-nine.vercel.app/about https://evalsuite-nine.vercel.app/release-notes https://evalsuite-nine.vercel.app/download https://evalsuite-nine.vercel.app/docs https://evalsuite-nine.vercel.app/docs/getting-started https://evalsuite-nine.vercel.app/docs/concepts https://evalsuite-nine.vercel.app/docs/metrics https://evalsuite-nine.vercel.app/docs/api https://evalsuite-nine.vercel.app/docs/cli https://evalsuite-nine.vercel.app/docs/classification https://evalsuite-nine.vercel.app/docs/regression https://evalsuite-nine.vercel.app/docs/clinical https://evalsuite-nine.vercel.app/docs/calibration https://evalsuite-nine.vercel.app/docs/statistics https://evalsuite-nine.vercel.app/docs/bootstrap https://evalsuite-nine.vercel.app/docs/model-comparison https://evalsuite-nine.vercel.app/docs/segmentation https://evalsuite-nine.vercel.app/docs/detection https://evalsuite-nine.vercel.app/docs/reporting https://evalsuite-nine.vercel.app/docs/llm https://evalsuite-nine.vercel.app/docs/text-generation https://evalsuite-nine.vercel.app/docs/factuality https://evalsuite-nine.vercel.app/docs/llm-judge https://evalsuite-nine.vercel.app/docs/rag https://evalsuite-nine.vercel.app/docs/structured-output https://evalsuite-nine.vercel.app/docs/llm-systems https://evalsuite-nine.vercel.app/docs/metrics/classification--accuracy https://evalsuite-nine.vercel.app/docs/metrics/classification--precision https://evalsuite-nine.vercel.app/docs/metrics/classification--recall https://evalsuite-nine.vercel.app/docs/metrics/classification--f1 https://evalsuite-nine.vercel.app/docs/metrics/classification--balanced_accuracy https://evalsuite-nine.vercel.app/docs/metrics/classification--mcc https://evalsuite-nine.vercel.app/docs/metrics/classification--cohen_kappa https://evalsuite-nine.vercel.app/docs/metrics/classification--roc_auc https://evalsuite-nine.vercel.app/docs/metrics/classification--pr_auc https://evalsuite-nine.vercel.app/docs/metrics/classification--log_loss https://evalsuite-nine.vercel.app/docs/metrics/classification--confusion_matrix https://evalsuite-nine.vercel.app/docs/metrics/regression--mae https://evalsuite-nine.vercel.app/docs/metrics/regression--mse https://evalsuite-nine.vercel.app/docs/metrics/regression--rmse https://evalsuite-nine.vercel.app/docs/metrics/regression--r2 https://evalsuite-nine.vercel.app/docs/metrics/regression--adjusted_r2 https://evalsuite-nine.vercel.app/docs/metrics/regression--mape https://evalsuite-nine.vercel.app/docs/metrics/regression--smape https://evalsuite-nine.vercel.app/docs/metrics/regression--rmsle https://evalsuite-nine.vercel.app/docs/metrics/regression--median_absolute_error https://evalsuite-nine.vercel.app/docs/metrics/regression--huber_loss https://evalsuite-nine.vercel.app/docs/metrics/regression--quantile_loss https://evalsuite-nine.vercel.app/docs/metrics/clinical--sensitivity https://evalsuite-nine.vercel.app/docs/metrics/clinical--specificity https://evalsuite-nine.vercel.app/docs/metrics/clinical--ppv https://evalsuite-nine.vercel.app/docs/metrics/clinical--npv https://evalsuite-nine.vercel.app/docs/metrics/clinical--lr_positive https://evalsuite-nine.vercel.app/docs/metrics/clinical--lr_negative https://evalsuite-nine.vercel.app/docs/metrics/clinical--diagnostic_odds_ratio https://evalsuite-nine.vercel.app/docs/metrics/clinical--youden_j https://evalsuite-nine.vercel.app/docs/metrics/clinical--net_benefit https://evalsuite-nine.vercel.app/docs/metrics/calibration--brier_score https://evalsuite-nine.vercel.app/docs/metrics/calibration--ece https://evalsuite-nine.vercel.app/docs/metrics/calibration--mce https://evalsuite-nine.vercel.app/docs/metrics/calibration--calibration_slope https://evalsuite-nine.vercel.app/docs/metrics/calibration--hosmer_lemeshow https://evalsuite-nine.vercel.app/docs/metrics/confidence--wilson https://evalsuite-nine.vercel.app/docs/metrics/confidence--clopper_pearson https://evalsuite-nine.vercel.app/docs/metrics/confidence--bootstrap_percentile https://evalsuite-nine.vercel.app/docs/metrics/confidence--bootstrap_bca https://evalsuite-nine.vercel.app/docs/metrics/statistics--t_test_independent https://evalsuite-nine.vercel.app/docs/metrics/statistics--t_test_paired https://evalsuite-nine.vercel.app/docs/metrics/statistics--mann_whitney https://evalsuite-nine.vercel.app/docs/metrics/statistics--wilcoxon https://evalsuite-nine.vercel.app/docs/metrics/statistics--kruskal_wallis https://evalsuite-nine.vercel.app/docs/metrics/statistics--friedman https://evalsuite-nine.vercel.app/docs/metrics/statistics--mcnemar https://evalsuite-nine.vercel.app/docs/metrics/statistics--shapiro_wilk https://evalsuite-nine.vercel.app/docs/metrics/statistics--delong https://evalsuite-nine.vercel.app/docs/metrics/effect-size--cohens_d https://evalsuite-nine.vercel.app/docs/metrics/effect-size--hedges_g https://evalsuite-nine.vercel.app/docs/metrics/effect-size--cramers_v https://evalsuite-nine.vercel.app/docs/metrics/multiple-testing--bonferroni https://evalsuite-nine.vercel.app/docs/metrics/multiple-testing--holm https://evalsuite-nine.vercel.app/docs/metrics/multiple-testing--hochberg https://evalsuite-nine.vercel.app/docs/metrics/multiple-testing--benjamini_hochberg https://evalsuite-nine.vercel.app/docs/metrics/segmentation--dice https://evalsuite-nine.vercel.app/docs/metrics/segmentation--iou https://evalsuite-nine.vercel.app/docs/metrics/segmentation--miou https://evalsuite-nine.vercel.app/docs/metrics/segmentation--pixel_accuracy https://evalsuite-nine.vercel.app/docs/metrics/segmentation--boundary_iou https://evalsuite-nine.vercel.app/docs/metrics/segmentation--hausdorff https://evalsuite-nine.vercel.app/docs/metrics/detection--box_iou https://evalsuite-nine.vercel.app/docs/metrics/detection--average_precision https://evalsuite-nine.vercel.app/docs/metrics/detection--map50_95 https://evalsuite-nine.vercel.app/docs/metrics/classification--fbeta https://evalsuite-nine.vercel.app/docs/metrics/classification--hamming_loss https://evalsuite-nine.vercel.app/docs/metrics/classification--jaccard https://evalsuite-nine.vercel.app/docs/metrics/classification--top_k_accuracy https://evalsuite-nine.vercel.app/docs/metrics/regression--explained_variance https://evalsuite-nine.vercel.app/docs/metrics/regression--max_error https://evalsuite-nine.vercel.app/docs/metrics/regression--mean_bias_error https://evalsuite-nine.vercel.app/docs/metrics/regression--msle https://evalsuite-nine.vercel.app/docs/metrics/regression--rae https://evalsuite-nine.vercel.app/docs/metrics/regression--rse https://evalsuite-nine.vercel.app/docs/metrics/calibration--calibration_intercept https://evalsuite-nine.vercel.app/docs/metrics/segmentation--mean_pixel_accuracy https://evalsuite-nine.vercel.app/docs/metrics/segmentation--average_surface_distance https://evalsuite-nine.vercel.app/docs/metrics/confidence--accuracy_ci https://evalsuite-nine.vercel.app/docs/metrics/confidence--roc_auc_ci https://evalsuite-nine.vercel.app/docs/metrics/statistics--chi_square_test https://evalsuite-nine.vercel.app/docs/metrics/statistics--fisher_exact_test https://evalsuite-nine.vercel.app/docs/metrics/effect-size--cliffs_delta https://evalsuite-nine.vercel.app/docs/metrics/statistics--paired_bootstrap_test https://evalsuite-nine.vercel.app/docs/metrics/statistics--compare https://evalsuite-nine.vercel.app/docs/metrics/factuality--abstention_accuracy https://evalsuite-nine.vercel.app/docs/metrics/factuality--answer_correctness https://evalsuite-nine.vercel.app/docs/metrics/factuality--answer_relevance https://evalsuite-nine.vercel.app/docs/metrics/factuality--citation_precision https://evalsuite-nine.vercel.app/docs/metrics/factuality--citation_recall https://evalsuite-nine.vercel.app/docs/metrics/factuality--claim_verification_accuracy https://evalsuite-nine.vercel.app/docs/metrics/factuality--exact_match https://evalsuite-nine.vercel.app/docs/metrics/factuality--faithfulness https://evalsuite-nine.vercel.app/docs/metrics/factuality--groundedness https://evalsuite-nine.vercel.app/docs/metrics/factuality--hallucination_rate https://evalsuite-nine.vercel.app/docs/metrics/factuality--knowledge_consistency https://evalsuite-nine.vercel.app/docs/metrics/factuality--token_f1 https://evalsuite-nine.vercel.app/docs/metrics/llm-judge--bradley_terry https://evalsuite-nine.vercel.app/docs/metrics/llm-judge--elo_ratings https://evalsuite-nine.vercel.app/docs/metrics/llm-judge--fleiss_kappa https://evalsuite-nine.vercel.app/docs/metrics/llm-judge--judge_agreement https://evalsuite-nine.vercel.app/docs/metrics/llm-judge--krippendorff_alpha https://evalsuite-nine.vercel.app/docs/metrics/llm-judge--win_rate https://evalsuite-nine.vercel.app/docs/metrics/llm-judge--position_consistency https://evalsuite-nine.vercel.app/docs/metrics/llm-judge--rubric_score https://evalsuite-nine.vercel.app/docs/metrics/llm-judge--self_preference_bias https://evalsuite-nine.vercel.app/docs/metrics/llm-judge--verbosity_bias https://evalsuite-nine.vercel.app/docs/metrics/rag--context_precision https://evalsuite-nine.vercel.app/docs/metrics/rag--context_recall https://evalsuite-nine.vercel.app/docs/metrics/rag--context_relevance https://evalsuite-nine.vercel.app/docs/metrics/rag--failure_attribution https://evalsuite-nine.vercel.app/docs/metrics/rag--hit_rate_at_k https://evalsuite-nine.vercel.app/docs/metrics/rag--latency_summary https://evalsuite-nine.vercel.app/docs/metrics/rag--mean_average_precision_at_k https://evalsuite-nine.vercel.app/docs/metrics/rag--mrr https://evalsuite-nine.vercel.app/docs/metrics/rag--ndcg_at_k https://evalsuite-nine.vercel.app/docs/metrics/rag--precision_at_k https://evalsuite-nine.vercel.app/docs/metrics/rag--recall_at_k https://evalsuite-nine.vercel.app/docs/metrics/rag--task_success_rate https://evalsuite-nine.vercel.app/docs/metrics/reasoning--benchmark_accuracy https://evalsuite-nine.vercel.app/docs/metrics/reasoning--majority_vote_accuracy https://evalsuite-nine.vercel.app/docs/metrics/reasoning--pass_at_k https://evalsuite-nine.vercel.app/docs/metrics/semantic--bertscore https://evalsuite-nine.vercel.app/docs/metrics/semantic--embedding_similarity https://evalsuite-nine.vercel.app/docs/metrics/semantic--model_score https://evalsuite-nine.vercel.app/docs/metrics/semantic--moverscore https://evalsuite-nine.vercel.app/docs/metrics/structured-output--api_call_success_rate https://evalsuite-nine.vercel.app/docs/metrics/structured-output--constraint_satisfaction_rate https://evalsuite-nine.vercel.app/docs/metrics/structured-output--format_compliance https://evalsuite-nine.vercel.app/docs/metrics/structured-output--instruction_compliance_rate https://evalsuite-nine.vercel.app/docs/metrics/structured-output--json_schema_compliance https://evalsuite-nine.vercel.app/docs/metrics/structured-output--json_validity https://evalsuite-nine.vercel.app/docs/metrics/structured-output--instruction_retention https://evalsuite-nine.vercel.app/docs/metrics/structured-output--required_field_accuracy https://evalsuite-nine.vercel.app/docs/metrics/structured-output--tool_argument_accuracy https://evalsuite-nine.vercel.app/docs/metrics/structured-output--tool_call_f1 https://evalsuite-nine.vercel.app/docs/metrics/structured-output--tool_selection_accuracy https://evalsuite-nine.vercel.app/docs/metrics/structured-output--extra_content_rate https://evalsuite-nine.vercel.app/docs/metrics/structured-output--xml_validity https://evalsuite-nine.vercel.app/docs/metrics/text-generation--bleu https://evalsuite-nine.vercel.app/docs/metrics/text-generation--chrf https://evalsuite-nine.vercel.app/docs/metrics/text-generation--cider https://evalsuite-nine.vercel.app/docs/metrics/text-generation--cross_entropy https://evalsuite-nine.vercel.app/docs/metrics/text-generation--distinct_n https://evalsuite-nine.vercel.app/docs/metrics/text-generation--mauve https://evalsuite-nine.vercel.app/docs/metrics/text-generation--meteor https://evalsuite-nine.vercel.app/docs/metrics/text-generation--perplexity https://evalsuite-nine.vercel.app/docs/metrics/text-generation--rouge_1 https://evalsuite-nine.vercel.app/docs/metrics/text-generation--rouge_2 https://evalsuite-nine.vercel.app/docs/metrics/text-generation--rouge_l https://evalsuite-nine.vercel.app/docs/metrics/text-generation--rouge_lsum https://evalsuite-nine.vercel.app/docs/metrics/text-generation--self_bleu https://evalsuite-nine.vercel.app/docs/metrics/text-generation--sentence_bleu https://evalsuite-nine.vercel.app/docs/metrics/text-generation--spice https://evalsuite-nine.vercel.app/docs/metrics/text-generation--ter https://evalsuite-nine.vercel.app/docs/metrics/agents--agent_cost_per_task https://evalsuite-nine.vercel.app/docs/metrics/agents--human_intervention_rate https://evalsuite-nine.vercel.app/docs/metrics/agents--invalid_tool_call_rate https://evalsuite-nine.vercel.app/docs/metrics/agents--plan_adherence https://evalsuite-nine.vercel.app/docs/metrics/agents--tool_failure_recovery_rate https://evalsuite-nine.vercel.app/docs/metrics/agents--state_tracking_accuracy https://evalsuite-nine.vercel.app/docs/metrics/agents--steps_per_task https://evalsuite-nine.vercel.app/docs/metrics/agents--task_completion_rate https://evalsuite-nine.vercel.app/docs/metrics/agents--tool_use_efficiency https://evalsuite-nine.vercel.app/docs/metrics/agents--unnecessary_tool_call_rate https://evalsuite-nine.vercel.app/docs/metrics/code--patch_acceptance_rate https://evalsuite-nine.vercel.app/docs/metrics/code--execution_success_rate https://evalsuite-nine.vercel.app/docs/metrics/code--codebleu https://evalsuite-nine.vercel.app/docs/metrics/code--syntax_validity_rate https://evalsuite-nine.vercel.app/docs/metrics/code--code_complexity https://evalsuite-nine.vercel.app/docs/metrics/code--coverage_rate https://evalsuite-nine.vercel.app/docs/metrics/code--resolved_rate https://evalsuite-nine.vercel.app/docs/metrics/code--runtime_efficiency https://evalsuite-nine.vercel.app/docs/metrics/code--security_vulnerability_rate https://evalsuite-nine.vercel.app/docs/metrics/code--static_analysis_violation_rate https://evalsuite-nine.vercel.app/docs/metrics/code--unit_test_pass_rate https://evalsuite-nine.vercel.app/docs/metrics/efficiency--inference_cost https://evalsuite-nine.vercel.app/docs/metrics/efficiency--energy_per_request https://evalsuite-nine.vercel.app/docs/metrics/efficiency--availability https://evalsuite-nine.vercel.app/docs/metrics/efficiency--token_usage https://evalsuite-nine.vercel.app/docs/metrics/efficiency--latency_percentiles https://evalsuite-nine.vercel.app/docs/metrics/efficiency--resource_utilization https://evalsuite-nine.vercel.app/docs/metrics/efficiency--requests_per_second https://evalsuite-nine.vercel.app/docs/metrics/efficiency--throughput https://evalsuite-nine.vercel.app/docs/metrics/efficiency--time_per_output_token https://evalsuite-nine.vercel.app/docs/metrics/efficiency--time_to_first_token https://evalsuite-nine.vercel.app/docs/metrics/long-context--citation_coverage https://evalsuite-nine.vercel.app/docs/metrics/long-context--context_utilization https://evalsuite-nine.vercel.app/docs/metrics/long-context--cross_document_consistency https://evalsuite-nine.vercel.app/docs/metrics/long-context--retrieval_accuracy_by_length https://evalsuite-nine.vercel.app/docs/metrics/long-context--lost_in_the_middle https://evalsuite-nine.vercel.app/docs/metrics/long-context--needle_in_haystack https://evalsuite-nine.vercel.app/docs/metrics/long-context--position_accuracy https://evalsuite-nine.vercel.app/docs/metrics/long-context--compression_ratio https://evalsuite-nine.vercel.app/docs/metrics/long-context--summary_coverage https://evalsuite-nine.vercel.app/docs/metrics/multilingual--code_switching_robustness https://evalsuite-nine.vercel.app/docs/metrics/multilingual--cross_lingual_consistency https://evalsuite-nine.vercel.app/docs/metrics/multilingual--cultural_appropriateness https://evalsuite-nine.vercel.app/docs/metrics/multilingual--language_consistency https://evalsuite-nine.vercel.app/docs/metrics/multilingual--language_id_accuracy https://evalsuite-nine.vercel.app/docs/metrics/multilingual--language_parity https://evalsuite-nine.vercel.app/docs/metrics/multilingual--bitext_mining_accuracy https://evalsuite-nine.vercel.app/docs/metrics/multilingual--direct_assessment https://evalsuite-nine.vercel.app/docs/metrics/robustness--adversarial_robustness https://evalsuite-nine.vercel.app/docs/metrics/robustness--contradiction_rate https://evalsuite-nine.vercel.app/docs/metrics/robustness--invariance_violation_rate https://evalsuite-nine.vercel.app/docs/metrics/robustness--distribution_shift_drop https://evalsuite-nine.vercel.app/docs/metrics/robustness--error_rate https://evalsuite-nine.vercel.app/docs/metrics/robustness--truncation_sensitivity https://evalsuite-nine.vercel.app/docs/metrics/robustness--ood_accuracy https://evalsuite-nine.vercel.app/docs/metrics/robustness--paraphrase_consistency https://evalsuite-nine.vercel.app/docs/metrics/robustness--prompt_sensitivity https://evalsuite-nine.vercel.app/docs/metrics/robustness--recovery_success_rate https://evalsuite-nine.vercel.app/docs/metrics/robustness--response_stability https://evalsuite-nine.vercel.app/docs/metrics/robustness--noise_robustness https://evalsuite-nine.vercel.app/docs/metrics/safety--attack_success_rate https://evalsuite-nine.vercel.app/docs/metrics/safety--harmful_response_rate https://evalsuite-nine.vercel.app/docs/metrics/safety--exposure https://evalsuite-nine.vercel.app/docs/metrics/safety--over_refusal_rate https://evalsuite-nine.vercel.app/docs/metrics/safety--pii_leakage_rate https://evalsuite-nine.vercel.app/docs/metrics/safety--policy_violation_rate https://evalsuite-nine.vercel.app/docs/metrics/safety--red_team_success_rate https://evalsuite-nine.vercel.app/docs/metrics/safety--refusal_rate https://evalsuite-nine.vercel.app/docs/metrics/safety--stereotype_preference https://evalsuite-nine.vercel.app/docs/metrics/safety--toxicity_score https://evalsuite-nine.vercel.app/docs/metrics/safety--weat_effect_size https://evalsuite-nine.vercel.app/docs/metrics/uncertainty--adaptive_calibration_error https://evalsuite-nine.vercel.app/docs/metrics/uncertainty--aurc https://evalsuite-nine.vercel.app/docs/metrics/uncertainty--confidence_accuracy_correlation https://evalsuite-nine.vercel.app/docs/metrics/uncertainty--coverage_at_risk https://evalsuite-nine.vercel.app/docs/metrics/uncertainty--risk_at_coverage https://evalsuite-nine.vercel.app/docs/metrics/uncertainty--selective_risk