AI Paper Atlas

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

PoeTone: A Framework for Constrained Generation of Structured Chinese Songci with LLMs

Zhan Qu, Shuzhou Yuan, Michael Färber

AAAI 2026 · Large language models › Instruction and fine-tuning · approach: Calibration and knowledge in LLM tuning · 2 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
Generate-Critic best-of-N SFT with rule-based criticQwen3-8B · Songci formal conformity (one-shot)Formal Conformity Score (%): +5.88Improved
Generate-Critic SFTQwen3-8B · Songci formal conformity (chain-of-thought)Formal Conformity Score (%): -3.11Hurt
Generate-Critic SFTQwen3-8B · Songci formal conformity (instruction)Formal Conformity Score (%): +0.33No clear effect
Generate-Critic SFTLlama-3.1-8B-Instruct · Songci formal conformity (all prompts)Formal Conformity Score (%): +1.11 to +1.63Improved
Generate-Critic SFTMistral-Small-Instruct-2409 · Songci formal conformity (one-shot)Formal Conformity Score (%): +3.40Improved
Generate-Critic SFTMistral-Small-Instruct-2409 · Songci formal conformity (instruction and CoT)Formal Conformity Score (%): -1.46 (instruction), -1.13 (CoT)Hurt
use chain-of-thought prompting instead of zero-shot (ablation)GPT-4o · Songci formal conformityFormal Conformity Score (%): 78.62 -> 71.11Hurt
use one-shot example instead of zero-shot (ablation)Llama-3.3-70B-Instruct · Songci formal conformityFormal Conformity Score (%): 35.88 -> 63.07Improved
use one-shot example instead of zero-shot (ablation)GPT-4o · Songci formal conformityFormal Conformity Score (%): 78.62 -> 76.59Hurt
use completion prompt instead of zero-shot (ablation)ERNIE 4.5 Turbo · Songci formal conformityFormal Conformity Score (%): 77.36 -> 82.38Improved
explicit rule instructions instead of one-shot (ablation)Qwen3-32B · Songci formal conformityFormal Conformity Score (%): 68.55 -> 56.51Hurt
higher formal conformity (Qwen3-32B) versus poetic qualityQwen3-32B · Songci poetic qualityhuman Overall Quality (1-5): 2.50 (vs DeepSeek-R1-0528-Qwen3-8B 2.85)Hurt

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