AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Selective Deferred Routing: Enabling Cost-Efficient Collaboration between Local SLMs and Remote LLMs

Qijun Miao, Zhixuan Fang

ICML 2026 · Large language models › Language, safety and interpretability · approach: LLM unlearning and benchmarks

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 SLM hidden states instead of BERT text embeddings (prompt only) (ablation)SDR MLP decider, Llama-3.1-8B · SQuADcost-performance AUC: 0.6013 -> 0.8978Improved
use SLM hidden states instead of BERT text embeddings (prompt only) (ablation)SDR MLP decider, Llama-3.1-8B · MMLUcost-performance AUC: 0.6047 -> 0.6509Improved
use SLM hidden states instead of BERT text embeddings (prompt only) (ablation)SDR MLP decider, Llama-3.1-8B · GSM8Kcost-performance AUC: 0.5476 -> 0.6706Improved
use SLM hidden states instead of BERT text embeddings (prompt only) (ablation)SDR MLP decider, Qwen-2.5-7B · SQuADcost-performance AUC: 0.6419 -> 0.9204Improved
add SLM answer to prompt input with text embeddings (ablation)SDR (P,T), Qwen-2.5-7B · GSM8Kcost-performance AUC: 0.4975 -> 0.7123Improved
add SLM answer to prompt input with text embeddings (ablation)SDR (P,T), Llama-3.1-8B · SQuADcost-performance AUC: 0.6013 -> 0.6590Improved
add SLM answer to prompt input with hidden states (ablation)SDR (P,H), Llama-3.1-8B · MMLUcost-performance AUC: 0.6509 -> 0.7008Improved
add SLM answer to prompt input with hidden states (ablation)SDR (P,H), Llama-3.1-8B · SQuADcost-performance AUC: 0.8978 -> 0.8946No clear effect
add SLM answer to prompt input with hidden states (ablation)SDR (P,H), Qwen-2.5-7B · SQuADcost-performance AUC: 0.9204 -> 0.9207No clear effect
add SLM answer to prompt input with hidden states (ablation)SDR (P,H), Qwen-2.5-7B · GSM8Kcost-performance AUC: 0.7461 -> 0.7746Improved
use full SDR (transformer-block decider) instead of best baseline FrugalGPTSDR, Llama-3.1-8B local · SQuAD (main results)cost-performance AUC: 0.5769 -> 0.9165Improved
use SDR instead of FrugalGPTSDR, Qwen-2.5-7B local · GSM8K (main results)cost-performance AUC: 0.7516 -> 0.7993Improved

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