AI Paper Atlas

Research map / Large language models / Instruction and fine-tuning

LLM DNA: Tracing Model Evolution via Functional Representations

Zhaomin Wu, Haodong Zhao, Ziyang Wang, Jizhou Guo, Qian Wang, Bingsheng He

ICLR 2026 · Large language models › Instruction and fine-tuning · approach: Fine-tuning generalization and delta compression · 9 citations (OpenAlex)

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
use DNA (BGE) + SVM instead of PhyloLM + SVM for relation detectionLLM DNA · LLM relation detection (305 LLMs)AUC: 0.788 -> 0.992Improved
replace EmbedLLM learned embeddings with frozen training-free DNAEmbedLLM matrix factorisation routing · Model routing (EmbedLLM dataset)Routing accuracy: 0.665 -> 0.672Improved
extract DNA from random word prompts instead of benchmark data (ablation)DNA (Qwen3) + SVM · LLM relation detectionAccuracy: 0.919 -> 0.949Improved
bypass chat template (completion-style prompts) for DNA extraction (ablation)DNA (Qwen3) · LLM relation detectionAccuracy: 0.919 -> 0.930Improved
bypass chat template for DNA extraction (ablation)DNA (MPNet) · LLM relation detectionAccuracy: 0.932 -> 0.958Improved
bypass chat template for DNA extraction (ablation)DNA (BGE) · LLM relation detectionAccuracy: 0.934 -> 0.945Improved
use a smaller sentence-embedding model (MPNet 0.1B, BGE 0.3B) instead of Qwen3-8B (ablation)DNA · DNA structure (Mantel test)Correlation: r>0.75, p<0.0001No clear effect
vary embedding output dimension p in {4,64,128,1024,4096} (ablation)DNA (Qwen3-8B-Embedding) · LLM relation detectionAccuracy/AUCNo clear effect
increase DNA dimension L from 4 up to 128 (ablation)DNA · LLM relation detectionAccuracy/AUCImproved
restrict to Qwen2Tokenizer models (harder same-tokenizer setting)DNA (MPNet) · LLM relation detectionAccuracy: 0.93 -> 0.86Hurt
same-tokenizer setting for PhyloLMPhyloLM · LLM relation detectionAccuracy: 0.742 -> 0.755Improved
vary model parameter size of test pairsDNA classifier · Size-vs-correctness analysisPoint-biserial correlation: -0.006 to 0.023, p 0.63 to 0.69No clear effect

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