Quantifying Positional Biases in Text Embedding Models
This addresses a problem for researchers and practitioners in Information Retrieval and semantic similarity by highlighting robustness issues in embedding models, though it is incremental as it quantifies an underexplored bias.
The study quantified positional biases in text embedding models, revealing that they disproportionately prioritize the beginning of inputs, with ablations at the start reducing cosine similarity by up to 12.3% more than at the end.
Embedding models are crucial for tasks in Information Retrieval (IR) and semantic similarity measurement, yet their handling of longer texts and associated positional biases remains underexplored. In this study, we investigate the impact of content position and input size on text embeddings. Our experiments reveal that embedding models, irrespective of their positional encoding mechanisms, disproportionately prioritize the beginning of an input. Ablation studies demonstrate that insertion of irrelevant text or removal at the start of a document reduces cosine similarity between altered and original embeddings by up to 12.3% more than ablations at the end. Regression analysis further confirms this bias, with sentence importance declining as position moves further from the start, even with with content-agnosticity. We hypothesize that this effect arises from pre-processing strategies and chosen positional encoding techniques. These findings quantify the sensitivity of retrieval systems and suggest a new lens towards embedding model robustness.