# Attention Calibration for Position-Fair Dense Retrieval

arXiv: https://arxiv.org/abs/2606.02737 · HTML: https://mlsanity.com/paper/2606.02737/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 15 references that are in the map.
- First posted on arXiv 1 Jun 2026.

## Builds on (most cited first, 8 of 8)

- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models](https://mlsanity.com/paper/2104.08663/) (2021)
- [M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation](https://mlsanity.com/paper/2402.03216/) (2024)
- [mGTE: Generalized Long-Context Text Representation and Reranking Models for Multilingual Text Retrieval](https://mlsanity.com/paper/2407.19669/) (2024)
- [Dense X Retrieval: What Retrieval Granularity Should We Use?](https://mlsanity.com/paper/2312.06648/) (2023)
- [An Empirical Study of Position Bias in Modern Information Retrieval](https://mlsanity.com/paper/2505.13950/) (2025)
- [PosIR: Position-Aware Heterogeneous Information Retrieval Benchmark](https://mlsanity.com/paper/2601.08363/) (2026)
- [Information Representation Fairness in Long-Document Embeddings: The Peculiar Interaction of Positional and Language Bias](https://mlsanity.com/paper/2601.16934/) (2026)

## Abstract (arXiv)

Dense retrieval compresses a passage into a single vector, but this compression is positionally skewed: early content dominates the embedding, and retrieval degrades when the relevant span appears later. Prior work proposed an inference-time method that counteracts this skew by equalizing the pooling token's attention across passage segments. However, (i) it redistributes attention at a fixed strength, (ii) it forces the pooling token's attention to itself to a fixed basket-level mass despite substantial variation across layers and architectures, and (iii) its effect on retrieval has not been evaluated. We introduce a strength coefficient that interpolates between uncalibrated and fully equalized attention, together with an efficient implementation that reduces peak calibration memory overhead from 5-7 GiB to under 1 MiB. Across three embedding models and two pooling schemes, moderate calibration provides a better retrieval trade-off than full equalization. We introduce a variant that preserves the pooling token's self-attention mass and redistributes only the remaining mass. On a position-aware retrieval benchmark spanning 10 languages and 31 domains, a configuration selected on English FineWeb-PosQ and transferred without tuning reduces position sensitivity in all 16 evaluated length-quartile, model, and retrieval-setting combinations, by up to 43% relative, while improving nDCG@10 by up to 4.8% relative and leaving general retrieval effectiveness on NanoBEIR essentially unchanged. Calibration runs at indexing time, adding no query-time latency. We release our code at github.com/impresso/fair-sentence-transformers