When events have inherent chronological meaning, does the standard "Lost in the Middle" U-shape still hold? Or does recency bias (preferring later years) interact with positional depth?
| # | Experiment | Setup | Hypothesis |
|---|---|---|---|
| 1 | Chronological vs Reverse vs Scrambled | Same events in chronological, reverse, or random order | Chronological shows weaker U-shape due to temporal scaffolding |
| 2 | Recency × Position Interaction | Year correlates with position (early=old, late=new) | Recency bias amplifies end-position advantage |
1pip install -r requirements.txt
2python run_all.py --model Qwen/Qwen2.5-1.5B-Instruct --num-events 100 --num-examples 50"Position Bias Index is 38% lower in chronological ordering (PBI=0.28) vs scrambled ordering (PBI=0.45, p<0.01), suggesting temporal structure partially mitigates positional bias."
1@software{temporal_position_bias,
2 title={Temporal Position Bias: How Chronological Ordering Affects Long-Context Retrieval},
3 author={abhshkp},
4 year={2026},
5 url={https://huggingface.co/abhshkp/temporal-position-bias}
6}