Downloads · 30 days
11
2% of all-time downloads
youngseok12/HyperCLOVA-X-SEED-Think-14B-success-consolidation-krc-r4min-ties
HyperCLOVA-X-SEED-Think-14B-success-consolidation-krc-r4min-ties is a text generation model from youngseok12. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as other.
This repository contains a standalone BF16 model derived from naver-hyperclovax/HyperCLOVAX-SEED-Think-14B. Three axis-specific LoRA adapters (K/R/C) were trained with a deliberately minimal capacity and merged with T…
Downloads · 30 days
11
2% of all-time downloads
All-time downloads
443
Public
Parameters
14.7B
29.5 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors29.5 GB · 100%
From the Hugging Face model README
This repository contains a standalone BF16 model derived from
naver-hyperclovax/HyperCLOVAX-SEED-Think-14B.
Three axis-specific LoRA adapters (K/R/C) were trained with a deliberately
minimal capacity and merged with TIES. The model name begins with
HyperCLOVA X as required by the base model license.
naver-hyperclovax/HyperCLOVAX-SEED-Think-14B9b74e35d4c7e4ffec489f4171273caca8948a2b9HyperCLOVAXForCausalLMsafetensors, standalone merged full model[1.0, 1.0, 1.0] for K/R/C, density 0.5Each axis LoRA was trained on rows where the pristine base model's own
correct reasoning trace (from a free-generation pass) was kept as the SFT
target, restructured answer-first (정답: X\n근거: ...). This is a
success-consolidation (STaR-family) construction: the base solves its own
training pool under three conditions (A: free generation, B1: answer-letter
only, C: choice-order shuffled), and only rows the base gets right are used.
| Axis | Benchmark target | AI Hub source | Arm | Rows |
|---|---|---|---|---|
| K | KMMLU-Pro | 71875 (필수의료 의학지식 QA) | correct_unstable | 300 |
| R | MuSR(Ko) | 71568 (경제·스포츠 숫자연산 MRC) | correct_unstable | 236 |
| C | Com2-main(Ko) | 71949 (인과관계 기반 추론) | correct_all | 300 |
correct_unstable = base answers correctly under free generation but fails
the letter-only or choice-shuffle condition (an unstable/fragile correct
answer). correct_all was used for C because its unstable pool was too
small (247 rows) to reach the target row count.
4, alpha: 8, dropout: 0.05q_proj, v_proj only5e-5, cosine scheduler, warmup ratio 0.031, effective batch size: 16 (per-device 4 × grad accum 4)4096, loss: assistant-token-only42This capacity (rank 4, 2 target modules) was chosen because prior K-AI leaderboard submissions on this base model showed every rank-4/2-module LoRA scoring higher than every rank-16/7-module LoRA, regardless of training data — the base model is already strong, and a larger LoRA perturbation degrades it more than task-specific data helps.
Public benchmark test data was not used for training.