Text Generation on NoveltyBench
10Diversityλ-PR
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| λ-PRSFT Dataset=UF, Backbone=MISTRAL-12B2026.04 | 10 | 1.1 | |
| λ-PRSFT Dataset=UF, Backbone=LLAMA-3.1-8B2026.04 | 10 | 1.1 | |
| λ-PRSFT Dataset=UF, Backbone=OLMO-2-13B2026.04 | 9.9 | 1.1 | |
| λ-PRSFT Dataset=UF, Backbone=PYTHIA-12B2026.04 | 9.9 | 1 | |
| λ-PRSFT Dataset=UF, Backbone=QWEN-3-8B2026.04 | 9.9 | 1.1 | |
| λ-PRSFT Dataset=UF, Backbone=PHI-4-14B2026.04 | 9.9 | 1.1 | |
| λ-PRSFT Dataset=ALPACA, Backbone=PHI-4-14B2026.04 | 9.6 | 1.2 | |
| λ-PRSFT Dataset=ALPACA, Backbone=MISTRAL-12B2026.04 | 9.5 | 1.3 | |
| λ-PRSFT Dataset=ALPACA, Backbone=LLAMA-3.1-8B2026.04 | 9.5 | 1.3 | |
| GEMSFT Dataset=UF, Backbone=PYTHIA-12B2026.04 | 9.5 | 1.4 | |
| TOFUSFT Dataset=UF, Backbone=PYTHIA-12B2026.04 | 9.5 | 1.4 | |
| λ-PRSFT Dataset=ALPACA, Backbone=OLMO-2-13B2026.04 | 9.4 | 1.3 | |
| λ-PRSFT Dataset=ALPACA, Backbone=PYTHIA-12B2026.04 | 9.3 | 1.1 | |
| FLSFT Dataset=UF, Backbone=PYTHIA-12B2026.04 | 9.3 | 1.6 | |
| TOFUSFT Dataset=UF, Backbone=LLAMA-3.1-8B2026.04 | 9.3 | 2.6 | |
| λ-PRSFT Dataset=ALPACA, Backbone=QWEN-3-8B2026.04 | 9.1 | 1.2 | |
| GEMSFT Dataset=UF, Backbone=OLMO-2-13B2026.04 | 9.1 | 2.8 | |
| GEMSFT Dataset=UF, Backbone=LLAMA-3.1-8B2026.04 | 9.1 | 2.6 | |
| TOFUSFT Dataset=UF, Backbone=MISTRAL-12B2026.04 | 9.1 | 2.6 | |
| TOFUSFT Dataset=UF, Backbone=OLMO-2-13B2026.04 | 9.1 | 2.9 | |
| GEMSFT Dataset=UF, Backbone=MISTRAL-12B2026.04 | 9 | 2.8 | |
| FLSFT Dataset=UF, Backbone=LLAMA-3.1-8B2026.04 | 8.9 | 2.8 | |
| TOFUSFT Dataset=UF, Backbone=PHI-4-14B2026.04 | 8.9 | 2 | |
| TOFUSFT Dataset=ALPACA, Backbone=PYTHIA-12B2026.04 | 8.8 | 3 | |
| FLSFT Dataset=UF, Backbone=OLMO-2-13B2026.04 | 8.8 | 3 | |
| GEMSFT Dataset=UF, Backbone=QWEN-3-8B2026.04 | 8.8 | 2.8 | |
| TOFUSFT Dataset=UF, Backbone=QWEN-3-8B2026.04 | 8.8 | 2.8 | |
| GEMSFT Dataset=ALPACA, Backbone=PYTHIA-12B2026.04 | 8.7 | 3 | |
| GEMSFT Dataset=UF, Backbone=PHI-4-14B2026.04 | 8.6 | 2 | |
| FLSFT Dataset=ALPACA, Backbone=PYTHIA-12B2026.04 | 8.5 | 3.2 | |
| CESFT Dataset=UF, Backbone=PYTHIA-12B2026.04 | 8.5 | 1.7 | |
| FLSFT Dataset=UF, Backbone=MISTRAL-12B2026.04 | 8.5 | 2.9 | |
| FLSFT Dataset=UF, Backbone=PHI-4-14B2026.04 | 8.4 | 2.3 | |
| TOFUSFT Dataset=ALPACA, Backbone=OLMO-2-13B2026.04 | 8.3 | 4.4 | |
| TOFUSFT Dataset=ALPACA, Backbone=LLAMA-3.1-8B2026.04 | 8.3 | 4.3 | |
| TOFUSFT Dataset=ALPACA, Backbone=MISTRAL-12B2026.04 | 8.2 | 4.4 | |
| FLSFT Dataset=UF, Backbone=QWEN-3-8B2026.04 | 8.2 | 2.9 | |
| GEMSFT Dataset=ALPACA, Backbone=OLMO-2-13B2026.04 | 8.1 | 4.6 | |
| GEMSFT Dataset=ALPACA, Backbone=LLAMA-3.1-8B2026.04 | 8.1 | 4.7 | |
| GEMSFT Dataset=ALPACA, Backbone=MISTRAL-12B2026.04 | 8 | 4.5 | |
| TOFUSFT Dataset=ALPACA, Backbone=QWEN-3-8B2026.04 | 7.9 | 4.2 | |
| CESFT Dataset=ALPACA, Backbone=PYTHIA-12B2026.04 | 7.7 | 3.2 | |
| FLSFT Dataset=ALPACA, Backbone=OLMO-2-13B2026.04 | 7.7 | 4.5 | |
| FLSFT Dataset=ALPACA, Backbone=LLAMA-3.1-8B2026.04 | 7.7 | 4.2 | |
| GEMSFT Dataset=ALPACA, Backbone=QWEN-3-8B2026.04 | 7.6 | 4.3 | |
| CESFT Dataset=UF, Backbone=LLAMA-3.1-8B2026.04 | 7.6 | 2.9 | |
| TOFUSFT Dataset=ALPACA, Backbone=PHI-4-14B2026.04 | 7.5 | 2 | |
| FLSFT Dataset=ALPACA, Backbone=MISTRAL-12B2026.04 | 7.4 | 4.3 | |
| FLSFT Dataset=ALPACA, Backbone=QWEN-3-8B2026.04 | 7.4 | 4.3 | |
| CESFT Dataset=UF, Backbone=MISTRAL-12B2026.04 | 7.3 | 3.2 | |
| CESFT Dataset=UF, Backbone=OLMO-2-13B2026.04 | 7.3 | 3 | |
| GEMSFT Dataset=ALPACA, Backbone=PHI-4-14B2026.04 | 7.2 | 1.8 | |
| FLSFT Dataset=ALPACA, Backbone=PHI-4-14B2026.04 | 7.1 | 2 | |
| CESFT Dataset=UF, Backbone=QWEN-3-8B2026.04 | 7 | 2.9 | |
| CESFT Dataset=UF, Backbone=PHI-4-14B2026.04 | 6.9 | 2.2 | |
| CESFT Dataset=ALPACA, Backbone=MISTRAL-12B2026.04 | 6.6 | 4.3 | |
| CESFT Dataset=ALPACA, Backbone=OLMO-2-13B2026.04 | 6.5 | 4.2 | |
| CESFT Dataset=ALPACA, Backbone=LLAMA-3.1-8B2026.04 | 6.5 | 4.2 | |
| CESFT Dataset=ALPACA, Backbone=QWEN-3-8B2026.04 | 6.5 | 4.2 | |
| CESFT Dataset=ALPACA, Backbone=PHI-4-14B2026.04 | 5.8 | 1.7 | |
| STATe of ThoughtsTemperature (T)=1.0, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 5.39 | 4.01 | |
| STATe of ThoughtsTemperature (T)=0.7, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 5.02 | 3.83 | |
| STATe of ThoughtsTemperature (T)=0.5, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 4.87 | 3.77 | |
| CoT w/ Action SpaceTemperature (T)=1.0, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 3.74 | 3.49 | |
| ToT w/ Action SpaceTemperature (T)=1.0, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 3.61 | 3.61 | |
| CoT w/ Action SpaceTemperature (T)=0.7, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 3.36 | 3.28 | |
| ToT w/ Action SpaceTemperature (T)=0.7, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 3.25 | 3.36 | |
| ToTTemperature (T)=1.0, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 3.16 | 3.32 | |
| ToT w/ Action SpaceTemperature (T)=0.5, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 3.16 | 3.29 | |
| CoT w/ Action SpaceTemperature (T)=0.5, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 3.07 | 3.15 | |
| ToTTemperature (T)=0.7, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 2.81 | 3.07 | |
| CoTTemperature (T)=1.0, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 2.8 | 3.07 | |
| ToTTemperature (T)=0.5, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 2.51 | 2.87 | |
| CoTTemperature (T)=0.7, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 2.44 | 2.85 | |
| I/O w/ Action SpaceTemperature (T)=1.0, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 2.44 | 2.71 | |
| CoTTemperature (T)=0.5, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 2.15 | 2.65 | |
| I/OTemperature (T)=1.0, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 2.01 | 2.64 | |
| I/O w/ Action SpaceTemperature (T)=0.7, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 2.01 | 2.41 | |
| I/O w/ Action SpaceTemperature (T)=0.5, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 1.74 | 2.23 | |
| I/OTemperature (T)=0.7, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 1.67 | 2.45 | |
| I/OTemperature (T)=0.5, Backbone=Qwen3-30B-A3B-Instruct2026.02 | 1.48 | 2.28 |