AI Paper Atlas

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

Context Tuning for In-Context Optimization

Jack Lu, Ryan Teehan, Zhenbang Yang, Mengye Ren

ICML 2026 · Large language models › Instruction and fine-tuning · approach: Efficient LLM fine-tuning systems · 0 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
initialize prefix from demonstration KV cache (CT-KV) instead of random-token Prefix TuningPrefix Tuning · BBH (Llama3-8B)accuracy (%): 52.7 -> 57.9Improved
initialize soft prompt from demonstrations (CT-Prompt) vs Prompt TuningPrompt Tuning · NLP-LR (GPT-2)accuracy (%): 41.4 -> 43.2Improved
increase Prompt/Prefix Tuning tokens m to match demonstration tokens (ablation)Prompt Tuning · BBH (Llama3-8B)accuracy (%): 50.8 -> 47.5Hurt
use KV-prefix CT-KV instead of CT-PromptContext Tuning · NLP-LR (GPT-2)accuracy (%): 43.2 -> 44.2Improved
apply CT-KV after TTTTTT · NLP-LR (GPT-2)accuracy (%): 44.1 -> 47.6Improved
add Leave-One-Out Masking (ablation)CT-KV · BBH (Llama3-8B)accuracy (%): 54.4 -> 57.9Improved
add Leave-One-Out Masking (ablation)CT-KV · ARC (Llama3.2-1B)accuracy (%): 23.8 -> 22.5Hurt
add Token Dropout (ablation)CT-KV · BBH (Llama3-8B)accuracy (%): 55.3 -> 57.9Improved
use CT-KV instead of ICL on large modelsICL · BBH (Qwen3-32B)accuracy (%): 63.1 -> 69.1Improved
use CT-KV instead of ICL with 100 demonstrations (many-shot)ICL · BBH 8 tasks (Llama3-8B-Instruct)accuracy (%): 46.5 -> 62.6Improved
train longer (64 vs 32 epochs) (ablation)Prefix Tuning · BBH (Llama3-8B)accuracy (%): 52.73 -> 50.49Hurt
train longer (16 to 64 epochs) (ablation)CT-KV · BBH (Llama3-8B)accuracy (%): 57.91 -> 57.88No 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