← 研踪 PaperTrailarXiv:2611.19356 · v1
A Comparative Evaluation of BERT-Based and LLM-Based Models for Indic Sentence Similarity
首次提交 2026-11-14 19:55
版本更新 2026-11-14 19:55
雷达观测 2026-11-17 12:06
摘要
The relative effectiveness of BERT-based encoders and large language model (LLM)-based embeddings for Indic semantic textual similarity remains insufficiently characterized. We compare eight BERT-based and five LLM-based models on the 1,500-pair STS Benchmark development set translated into ten Indic languages and on the native Marathi MahaSTS test set. IndicSBERT achieves the highest Pearson correlation in all ten translated language sets, with a macro-average of 0.843, followed by BGE-M3 at 0.792. Harrier-270M is the strongest LLM-based model on translated STS-B, with a macro-average of 0.751. On MahaSTS, IndicSBERT and Bhasha-Embed-v0 achieve Pearson correlations of 0.901 and 0.900, respectively, while EmbeddingGemma leads the LLM-based models at 0.871. The highest-scoring BERT-based model thus exceeds the highest-scoring LLM-based model in both settings, although the magnitude of the difference varies across benchmarks. These findings indicate that LLM-derived embeddings do not uniformly improve Indic sentence similarity over established encoders and motivate evaluation that accounts for language coverage, task supervision, and benchmark construction.
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-14 19:55 · 最新观测版本
A Comparative Evaluation of BERT-Based and LLM-Based Models for Indic Sentence Similarity
The relative effectiveness of BERT-based encoders and large language model (LLM)-based embeddings for Indic semantic textual similarity remains insufficiently characterized. We compare eight BERT-based and five LLM-based models on the 1,500-pair STS Benchmark development set translated into ten Indic languages and on the native Marathi MahaSTS test set. IndicSBERT achieves the highest Pearson correlation in all ten translated language sets, with a macro-average of 0.843, followed by BGE-M3 at 0.792. Harrier-270M is the strongest LLM-based model on translated STS-B, with a macro-average of 0.751. On MahaSTS, IndicSBERT and Bhasha-Embed-v0 achieve Pearson correlations of 0.901 and 0.900, respectively, while EmbeddingGemma leads the LLM-based models at 0.871. The highest-scoring BERT-based model thus exceeds the highest-scoring LLM-based model in both settings, although the magnitude of the difference varies across benchmarks. These findings indicate that LLM-derived embeddings do not uniformly improve Indic sentence similarity over established encoders and motivate evaluation that accounts for language coverage, task supervision, and benchmark construction.
查看此版本解读 →查看此版本原文 ↗