AI Paper Atlas

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

Flora: Effortless Context Construction to Arbitrary Length and Scale

Tianxiang Chen, Zhentao Tan, Xiaofan Bo, Yue Wu, Tao Gong, Qi Chu, Jieping Ye

AAAI 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
fine-tune on Flora-80k long-context dataLlama-3-8B-Instruct · LongBench v1average score: 38.93 -> 50.73Improved
fine-tune on Flora-128k long-context dataQwQ-32B · LongBench v1average score: 66.23 -> 70.07Improved
fine-tune on Flora-128kQwQ-32B · LongBench v2 (CoT)average accuracy: 45.6 -> 50.5Improved
fine-tune on LongAlpaca-12k instead of FloraQwQ-32B · LongBench v2 (CoT)average accuracy: 45.6 -> 42.5Hurt
add Fewshot QA strategy (1/4 of data) (ablation)Mosaic-IT baseline · LongBench v1average score: 45.54 -> 47.40Improved
add Answer Before or After strategy (1/4 of data) (ablation)Mosaic-IT baseline · LongBench v1average score: 45.54 -> 47.93Improved
add Answer No Answer strategy (1/4 of data) (ablation)Mosaic-IT baseline · LongBench v1average score: 45.54 -> 47.97Improved
add Answer to ID strategy (1/4 of data) (ablation)Mosaic-IT baseline · LongBench v1average score: 45.54 -> 48.38Improved
combine all four strategies with Mosaic (3/7 base + 4/7 all) (ablation)Mosaic-IT baseline · LongBench v1average score: 45.54 -> 48.99Improved
replace augmented samples under 2k tokens with original SFT samples (ablation)Flora 3/7+4/7 mix · Short-context tasksaverage score: 43.02 -> 43.41Improved
use proposed exponential token-length distribution instead of normal (ablation)Flora 8k-sample set · LongBench v2average accuracy: 30.5 -> 32.3Improved
use proposed distribution instead of normal (ablation)Flora 8k-sample set · Short-context tasksaverage score: 43.56 -> 44.82Improved

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