AI Paper Atlas

Research map / Probabilistic and scientific modeling / Dynamical systems

The Hippocampal Place Field Gradient: A Bio-inspired Framework Building Multiscale Representation for Better Sample Efficiency

ZHOU Shujun, Junrong Qi, Guozhang Chen

ICML 2026 · Probabilistic and scientific modeling › Dynamical systems · approach: Synaptic plasticity learning rules

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace RoPE with HIPETransformer · One-to-One associative recallaccuracy (%): 81.56 -> 88.53Improved
replace RoPE with HIPETransformer · Block retrievalaccuracy (%): 96.32 -> 97.10Improved
use HIPE bipartition instead of RoPE20M Transformer · WikiText-103 L=2048perplexity: 133.81 -> 132.97Improved
use HIPE bipartition instead of RoPE60M Transformer · C4 L=1024perplexity: 141.65 -> 130.47Improved
use HIPE instead of RoPE at larger scale300M Transformer · C4 1B tokensperplexity: 28.93 -> 28.97No clear effect
use HIPE instead of RoPE with few labels300M LoRA · SST-2 100 samplesaccuracy (%): 67.3 -> 68.5Improved
use HIPE instead of RoPE with full training set300M LoRA · SST-2 full dataaccuracy (%): 82.03 -> 81.77Hurt
add HIPE (sigma_p=500) to deep layers with local-global attention (ablation)Local-Global · C4 60M L=1024perplexity: 59.36 -> 59.01Improved
add HIPE to deep layers of global-attention baseline (ablation)Baseline · C4 60M L=1024perplexity: 62.82 -> 62.71Improved
use uniform sigma_p=50 across all layers instead of bipartition (ablation)HIPE · WikiText-103 20M L=2048perplexity: 133.81 -> 138.45Hurt
use reverse schedule (broad shallow, sharp deep) instead of sharp-to-broad (ablation)HIPE · WikiText-103 20M L=2048perplexity: 136.22 vs 132.97Hurt
use continuous gradient instead of discrete bipartite schedule (ablation)HIPE · WikiText-103 20M L=2048perplexity: +2.56Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy