Youngjoon Jang

yjoonjang

·

https://yjoonjang.github.io/

AI & ML interests

Information Retrieval (IR), Retrieval-Augmented Generation (RAG)

Recent Activity

upvoted an article 3 days ago

mDenseOn with the mLateOn: Open Multilingual, Long-Context, and Code Retrieval Models

upvoted a collection 3 days ago

mDenseOn & mLateOn

upvoted a paper 4 days ago

SHIFT: Semantic Harmonization via Index-side Feature Transformation for Multilingual Information Retrieval

View all activity

Organizations

New activity in whybe-choi/ko-vdr-hn 20 days ago

Update README with old README

#15 opened 20 days ago by

New activity in whybe-choi/ko-vdr-hn 21 days ago

Restore dataset_info for base configs (corpus/naive/marginpos/percpos/shifted_by_n)

#14 opened 21 days ago by

Add distill_hardest + ensembledistill_{stratified,hardest} subsets

#13 opened 21 days ago by

Add ensembledistill_hardest subset

#12 opened 21 days ago by

Add ensembledistill_stratified subset

#11 opened 21 days ago by

Add distill_hardest subset

#10 opened 21 days ago by

Add distill_stratified subset

#9 opened 21 days ago by

New activity in whybe-choi/ko-vdr-hn 24 days ago

Add reranked_hardest config & rename reranked -> reranked_stratified

#8 opened 24 days ago by

New activity in whybe-choi/ko-vdr-hn 28 days ago

Add ko-vdr-train-public-v2.0_reranked config (KL-distillation set)

#7 opened 28 days ago by

Add ko-vdr-train-public-v2.0_reranked config (KL-distillation set)

#6 opened 28 days ago by

New activity in whybe-choi/ko-vdr-hn about 1 month ago

Replace ko-vdr-train-public-v2.0_reranked with training-ready flat schema (positive + 8 negatives + reranked_scores)

#5 opened about 1 month ago by

Rename reranker config to ko-vdr-train-public-v2.0_reranked + note source coverage

#4 opened about 1 month ago by

Add ko-vdr-train-public-v2.0_reranked config (Qwen3-VL reranker-scored candidates, mined from ko-vdr-train-public-v2.0 only)

#3 opened about 1 month ago by

Add reranked_qwen3_vl_8b config (Qwen3-VL reranker-scored candidates for distillation)

#2 opened about 1 month ago by

New activity in google/embeddinggemma-300m 6 months ago

Question regarding the two-layer linear (FNN) projection structure in EmbeddingGemma-300M

#33 opened 8 months ago by

New activity in yjoonjang/inference-free-splade-ko-v1 6 months ago

안녕하세요.

#1 opened 6 months ago by

New activity in skt/A.X-Encoder-base 6 months ago

Set tokenizer "model_max_length" property to 16384

#1 opened 6 months ago by

New activity in yjoonjang/colbert-ko-v1 9 months ago

Produce ColBERT-KO Evaluation Results

#1 opened 10 months ago by

New activity in lightonai/embeddings_supervised 10 months ago

How are the negatives mined?

#2 opened 10 months ago by

New activity in google/embeddinggemma-300m 11 months ago

Very bad performances (not gpu time, score)

#3 opened 11 months ago by