Qwen3-1.7B
A small-parameter language model in Alibaba's Qwen series, equipped with fundamental language generation and analysis capabilities. Designed for lightweight research and inference tasks, it balances compact model size with general-purpose language understanding.
Timeline (last 90 days)
Qwen3-1.7B's overall causal faithfulness is only 54.8%, far below that of Qwen3-4B.
Qwen3-1.7B's faithfulness is 68% on 2-hop tasks but plummets to 30% on 6-hop tasks, whereas Qwen3-4B remains relatively stable across different hop counts.
The author used the Qwen3-1.7B model for both generation and attribution analysis tasks.
All Facts (3)
Qwen3-1.7B's overall causal faithfulness is only 54.8%, far below that of Qwen3-4B.
50%UnverifiedQwen3-1.7B's faithfulness is 68% on 2-hop tasks but plummets to 30% on 6-hop tasks, whereas Qwen3-4B remains relatively stable across different hop counts.
50%UnverifiedThe author used the Qwen3-1.7B model for both generation and attribution analysis tasks.
50%