Table 8 Impact of Proposer Agent combination diversity on the performance of SLM-MATRIX

From: SLM-MATRIX: a multi-agent trajectory reasoning and verification framework for enhancing language models in materials data extraction

Proposer Agent Combination Strategy

Specific Model Combination

Accuracy(%)

Benchmark Heterogeneous Combination

Qwen2.5-7B-Instruct-Turbo, Mistral-7B-Instruct, gemma-2-9b-it

92.85% (±2.05%)

Low-Diversity Combination

Qwen2.5-7B-Instruct-Turbo

80.45% (±2.23%)

Alternative Heterogeneous Combination

Qwen2.5-Coder-7B, Mistral-7B-Instruct, gemma-2-9b-it

90.88% (±1.16%)

  1. All experiments were conducted with three Proposer Agents in each setting.
Back to article page