Model-Specific Task Similarity for Vision–Language Model Selection via Layer Conductance
Wei Yang, Hong Xie, Tao Tan, Xin Li, Defu Lian, Enhong Chen · 2026
Read Model-Specific Task Similarity for Vision–Language Model Selection via Layer Conductance by Wei Yang, Hong Xie, Tao Tan, Xin Li, Defu Lian, Enhong Chen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-01346.txt.
Description
Wei Yang and co-authors pick the best vision-language net for a job from inner workings, not brute tests. Open vision-language nets swarm, yet full tests cost too much, above all few-shot. Past pick plans lean on data-hungry proxies or flat tags missing one-way, model-bound transfer. Their plan reads each job via layer-wise conductance and builds a target-tuned block map via entropy alignment. A directional conductance gap score then forecasts target ranks from sources, beating top baselines by 14.7% on NDCG@5.
Catalog record
| CREATOR | Wei Yang, Hong Xie, Tao Tan, Xin Li, Defu Lian, Enhong Chen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 59683 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602.01346