Josiefied-Qwen2.5-14B-Instruct-abliterated-v4: benchmark results

Provider: Other. Released 2024-10-21. Access: Open.

Unified ELO 1618 ± 58, rank #560 of 2656 rated models, from 11 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Open LLM Leaderboard - IFEval82.92Score99.2
Open LLM Leaderboard - MATH Level 554.23Score98.9
Open Portuguese LLM - ASSIN2 RTE94.36Macro F1 (%)96
Open Portuguese LLM - OAB Exams61.41Accuracy (%)93.7
Open Portuguese LLM - BLUEX73.16Accuracy (%)93.5
Open Portuguese LLM - ENEM80.06Accuracy (%)93.3
Open LLM Leaderboard - MMLU-Pro50.18Score87.8
Open LLM Leaderboard - BBH63.56Score87.6
Open LLM Leaderboard - GPQA34.23Score85.1
Open Portuguese LLM - FaQuAD NLI78.11Macro F1 (%)82.7
Open LLM Leaderboard - MuSR42.87Score69.2

Interactive version: theaggregate.ai/model?slug=josiefied-qwen2-5-14b-instruct-abliterated-v4 · How It Works · Data refreshed daily, snapshot 2026-09-19.