Imprint

Follow this topic Ask about this topic

Momentum

2 papers in the last four weeks, with none the four weeks before. 0.0% of all new papers.

Jul 13Week of Sep 28

Latest papers 14

All topics

Imprint Reader: From Weight-Update Readout to Behavioral Intervention
Sep 28, 2026Guanxu Chen, Qihao Lin, Jing ShaoLatent Feature Interventions Readout
Shanghai Artificial Intelligence Laboratory · Shanghai Jiao Tong University
Story Imprinting: AI Assistants Absorb Traits from Human Characters They Resemble
Sep 12, 2026Jorio Cocola, Lev McKinney, Harry Mayne +2Artificial Intelligence Personas Artificial Intelligence Assistants
Truthful AI · Harvard University · METR +1
IMPRINT: Image-Conditioned Query Enrichment for Long-Tail Object Goal Navigation
Jul 27, 2026Jelin Raphael Akkara, Filippo Ziliotto, Luciano Serafini +2Vision-Language Navigation Semantic Mapping
University of Padova, Italy. · Fondazione Bruno Kessler (FBK), Trento, Italy.
Inverse-IMPRESSION: A Graph-based Platform for Molecular Structure Elucidation from Experimental NMR Spectroscopic Properties
Jul 10, 2026Zheqi Jin, Grace Armitage, Richard Cox +3Molecules Imprint
School of Chemistry, University of Bristol, Cantock’s Close, Bristol, BS8 1TS, U.K. · School of Engineering Mathematics and Technology, University of Bristol, Ada Lovelace Building, Tankard's Cl, Bristol BS8 1TW
Time Imprint: Learning Time-Aware Representations in Multi-Modal Knowledge Graphs
Jul 8, 2026Pengyu Zhang, Klim Zaporojets, Congfeng Cao +2Knowledge Graph Embeddings Multimodal Representations
University of Amsterdam Amsterdam, The Netherlands · Aarhus University Aarhus, Denmark
Rating the Pitch, Not the Product: User Evaluations of LLMs Reflect Expectations More Than Performance
Jul 6, 2026Robert Morabito, Tyler McDonald, Charitra Viswanath +4Large Language Model Evaluation Review
Brock University · Emory University · University of Southern California +2
Imprint: Online Memory Compression for Long-Horizon Egocentric QA
Jul 1, 2026Kousik Das, Debaditya RoyImprint Data Compression Methods
IIT, Kharagpur India
Transition-Aware best-of-N sampling for Longitudinal Chest X-ray Reports
Jun 23, 2026Halil Ibrahim Gulluk, Max Van Puyvelde, Wim Van Criekinge +1Chest X-Ray Radiology Report Generation
Department of Electrical Engineering, Stanford University, Stanford, CA, USA · Department of Biomedical Data Science, Stanford University School of Medicine, Stanford, CA, USA · Department of Mathematical Modelling, Statistics & Bioinformatics, Ghent University, Ghent, Belgium
HMAF: A Hierarchical Multi-Slot GD-RTB Allocation Framework
Jun 4, 2026Tianxing Bu, Zhaoqi Zhang, Linyou Cai +7Ads Power Allocation
Meituan Beijing, China · Nanyang Technological University Singapore · China Agricultural University Beijing, China
Deep Learning-based 3D Oral Cavity Reconstruction Using 2D Intraoral Images
Jun 4, 2026Jihun Cho, Soo-Yeon Jeong, Eun-Jeong Bae +1Intraoral Scans Occlusal Constraint
Dept. of Computer Engineering Pai Chai University Daejeon, Korea · Division of Software Engineering Pai Chai University Daejeon, Korea · Dept. of Dental Technology Bucheon University Bucheon, Korea
Perception or Prejudice: Can MLLMs Go Beyond First Impressions of Personality?
May 21, 2026Caixin Kang, Tianyu Yan, Sitong Gong +8Personality Multimodal Large Language Models
1The University of Tokyo · 2Shanda AI Research Tokyo · 3Dalian University of Technology
PGID: Progressive Guided Inversion and Denoising for Robust Watermark Detection
May 10, 2026Minh Quoc Duong, Chun Tong Lei, Chun Pong LauWatermarks Multi-Bit Watermarking
City University of Hong Kong
Improving Graph Few-shot Learning with Hyperbolic Space and Denoising Diffusion
Apr 30, 2026Yonghao Liu, Jialu Sun, Wei Pang +4Graph Representation Learning Few-Shot Learning
Key Laboratory of Symbolic Computation and Knowledge Engineering of the Ministry of Education, College of Computer Science and Technology, Jilin University · School of Mathematical and Computer Sciences, Heriot-Watt University · Department of Information Engineering and Computer Science, University of Trento
Platonic Representations for Poverty Mapping: Unified Vision-Language Codes or Agent-Induced Novelty?
Aug 1, 2025Satiyabooshan Murugaboopathy, Connor T. Jerzak, Adel DaoudMultimodal Dataset Vision-Language Foundation Models
AI and Global Development Lab, aidevlab.org · Institute for Analytical Sociology, Linköping University, Sweden · Institute of Computer Science, University Leipzig, Germany +1

Sep 28, 2026, cs.AIJ/K move · Enter open · S save

Imprint Reader: From Weight-Update Readout to Behavioral Intervention

Guanxu Chen, Qihao Lin, Jing Shao

Shanghai Artificial Intelligence Laboratory · Shanghai Jiao Tong University

Latent Feature Interventions Readout Imprint Data Generation Weight Decay Mathematical Reasoning

Open paper AskPDF arXiv

As language models take a growing role in AI development, a natural aspiration is for them to reflect on their own learning process, as humans do, and use that reflection to improve themselves. At the same time, these models have an advantage that human learners lack, since training leaves parameter-level traces that can, in principle, be inspected directly. However, current models cannot decode these traces into an explicit account of what they have learned. To this end, we introduce the \textit{Imprint Reader}, a model trained with \textit{Semantic Mount-and-Read Tuning} (SaRT) to describe frozen weight updates. SMaRT mounts each update onto the Reader and uses an anchor-free meta-query to elicit a natural-language description, while no-change and random-perturbation controls discourage unsupported claims. On held-out updates, the joint Reader reaches judge-based Pass@100 of

2\%

for knowledge and

16\%

for behavior. These results demonstrate the feasibility of natural-language readout while pointing to reliability across updates as the next step. Beyond free-form generation, the Reader provides a differentiable proxy for the gap between a specified target behavior and a candidate weight update. Its coordinate-aligned gradients support intervention through MetaEdit. At a

0.5\%

pruning rate, Reader-guided selection raises measured harmful-prompt refusal from

57.9\%

64.1\%

under a safety-maintenance target. Using behavior descriptions without target-task training data, MetaEdit increases the frequency of backtracking and sub-goal expressions in mathematical reasoning traces and raises BFCL Overall from

41.69\%

44.60\%

Momentum

Related topics

Latest papers 14