benchmark

CCL25-Eval Task 5

benchmarkactiveprovisionalccl25-eval-task-5-c9c098d3·1 events·first seen 6d ago

Aliases: CCL25-Eval Task 5

Co-occurring entities

CCPoetry-49K Qwen2.5-14B LoRA (Low-Rank Adaptation)PoetryQwen

More like this (12)

Claw-Eval InstructS2S-Eval T-Eval ProActEval Codex HumanEval GIFT-Eval KSAA-2026 Shared Task Cranfield evaluation paradigm BenHalluEval Cybersecurity Task Evaluation LAMDA-CL paired-scenario forced-choice probe

Recent events (1)

3arXiv · cs.CL·6d ago·source ↗

PoetryQwen: LoRA-fine-tuned Qwen2.5-14B for classical Chinese poetry understanding with new 49K dataset

Researchers introduce CCPoetry-49K, a 49,404-pair instruction dataset for classical Chinese poetry appreciation, decomposed into term interpretation, semantic interpretation, and emotional inference subtasks. They fine-tune Qwen2.5-14B using LoRA to produce PoetryQwen, achieving a 9.7% improvement over the baseline on the CCL25-Eval Task 5 benchmark (0.757 vs 0.690). The work addresses a gap in domain-specific LLM adaptation for classical Chinese literary tasks.

Open Weights Progress CCPoetry-49K Qwen2.5-14B LoRA (Low-Rank Adaptation)+2 more