Natural Language Understanding on EuroEval English
56.8EuroEval ScoreEuroLLM-9B (Base)
Evaluation Results
| Method | Links | |
|---|---|---|
| EuroLLM-9B (Base)Language model=EuroLLM-9B2026.05 | 56.8 | |
| aya-3B (Base)Language model=aya-3B2026.05 | 54.1 | |
| aya-3B (Multilingual All Lang SFT)Regime=Multilingual, Post-training=SFT, Data=All Lang (Dolci SFT)2026.05 | 7.9 | |
| aya-3B (Multilingual Max-R SFT)Regime=Multilingual, Post-training=SFT, Data=Max-R2026.05 | 7 | |
| aya-3B (Monolingual Max-R SFT)Regime=Monolingual, Post-training=SFT, Data=Max-R2026.05 | 6.3 | |
| EuroLLM-9B (Multilingual All Lang SFT)Regime=Multilingual, Post-training=SFT, Data=All Lang (Dolci SFT)2026.05 | 3.9 | |
| ICR (8B)Parameters=8B, Comparison=EuroLLM-9B2026.05 | 3.6 | |
| aya-3B (Monolingual In-lang SFT)Regime=Monolingual, Post-training=SFT, Data=In-lang (Dolci SFT)2026.05 | 2.8 | |
| MAPO (13B)Parameters=13B, Comparison=EuroLLM-9B2026.05 | 2.6 | |
| ICR (8B)Baseline Model=ICR, Parameters=8B, Comparison=aya-3B2026.05 | 1 | |
| EuroLLM-9B (Monolingual Max-R SFT)Regime=Monolingual, Post-training=SFT, Data=Max-R2026.05 | 0.5 | |
| aya-3B (Multilingual Paired DPO)Regime=Multilingual, Post-training=DPO, Data=Paired2026.05 | 0.4 | |
| EuroLLM-9B (Monolingual Paired DPO)Regime=Monolingual, Post-training=DPO, Data=Paired2026.05 | 0.4 | |
| EuroLLM-9B (Multilingual Paired DPO)Regime=Multilingual, Post-training=DPO, Data=Paired2026.05 | 0.3 | |
| aya-3B (Monolingual Paired DPO)Regime=Monolingual, Post-training=DPO, Data=Paired2026.05 | 0.2 | |
| EuroLLM-9B (Multilingual Max-R SFT)Regime=Multilingual, Post-training=SFT, Data=Max-R2026.05 | 0.2 | |
| EuroLLM-9B (Monolingual In-lang SFT)Regime=Monolingual, Post-training=SFT, Data=In-lang (Dolci SFT)2026.05 | 0.1 | |
| MAPO (13B)Baseline Model=MAPO, Parameters=13B, Comparison=aya-3B2026.05 | 0 |