List your company

Question Answering on HotpotQA (EM, F1, R-L, BLEU)

50.1ROUGE-LLLaDA-8B + OSCAR
1.906414.418226.9339.4418Apr 2, 2026Apr 10, 2026Apr 19, 2026Apr 28, 2026May 6, 2026May 15, 2026May 24, 2026

Evaluation results

MethodLinks
LLaDA-8B + OSCAROSCAR correction=After2026.0450.143.652.730.1
Dream-7B + OSCAROSCAR correction=After2026.0443.737.446.1—
LLaDA-8BOSCAR correction=Before2026.0442.83745.224.6
HMTModel=Qwen2.5-14B2026.0542.4———
Dream-7BOSCAR correction=Before2026.0436.530.238.8—
HMTModel=Llama3.1-8B2026.0530.54———
H2MTModel=Qwen2.5-14B2026.0530.37———
H2MTModel=Qwen3-4B2026.0530.37———
HMTModel=Qwen2.5-7B2026.0530.06———
H2MTModel=Qwen2.5-7B2026.0528.89———
HMTModel=Qwen3-4B2026.0528.21———
HMTModel=SmolLM-135M2026.0527.44———
H2MTModel=Llama3.1-8B2026.0525.37———
H2MTModel=SmolLM-135M2026.0523.8———
BackboneModel=Llama3.1-8B2026.0519.99———
BackboneModel=Qwen2.5-7B2026.0519.6———
BackboneModel=Qwen2.5-14B2026.0519———
BackboneModel=Qwen3-4B2026.0518.21———
BackboneModel=SmolLM-135M2026.053.76———