Research map / Large language models / Agents, code and math
Type-Compliant Adaptation Cascades
Chu-Cheng Lin, Daiyi Peng, Yifeng Lu, Ming Zhang, Eugene Ie
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 changed | Where | Measured | Effect |
|---|---|---|---|
| adapt weights with TACSTaR instead of using un-adapted base weights (ablation) | cot-cascade-structure · MGSM | accuracy: 45.4 -> 82.2 | Improved |
| adapt weights with TACSTaR instead of un-adapted (ablation) | cot-cascade-structure · FinQA | accuracy: 13.0 -> 36.0 | Improved |
| adapt weights with TACSTaR instead of un-adapted (ablation) | refine-structure · HotPotQA | accuracy: 24.0 -> 39.0 | Improved |
| use typed TACSTaR instead of untyped STaR | TAC · MGSM (gemma-1.1-7b-it) | accuracy: 10.5% -> 27.3% | Improved |
| use TAC instead of best DSPy on structured output | TAC · MGSM-SymPy | accuracy: 57.1% -> 75.9% | Improved |
| use TAC instead of best DSPy on small model | TAC · FinQA (gemma-1.1-7b-it) | accuracy: 0.7% -> 9.7% | Improved |
| add amortized inference network at training time (ablation) | TACSTaR · FinQA | accuracy: 36.0 -> 41.7 | Improved |
| add amortized inference network at training time (ablation) | TACSTaR · MGSM | accuracy: 82.2 -> 82.4 | No clear effect |
| use renormalized posterior classification instead of unconstrained generation (ablation) | TAC · MuSR (gemma-2-27b-it) | accuracy: 51.6 -> 65.0 | Improved |
| decouple rationale and output with separate adaptors (cascade vs type structure) (ablation) | TAC · MGSM | accuracy: 80.4 -> 82.2 | Improved |
| use CoT cascade structure instead of direct | TAC · HotPotQA | accuracy: 34.0 -> 32.0 | Hurt |
| use refine-structure instead of direct | TAC · HotPotQA | accuracy: 34.0 -> 39.0 | Improved |
Closest papers by meaning
- Unleashing the Potential of Large Language Models as Prompt Optimizers: Analogical Analysis with Gradient-based Model Optimizers (AAAI 2025 · similarity 0.87)
- ELICIT: LLM Augmentation Via External In-context Capability (ICLR 2025 · similarity 0.87)
- Test-Time Learning for Large Language Models (ICML 2025 · similarity 0.87)
- Unveiling the Secret Recipe: A Guide For Supervised Fine-Tuning Small LLMs (ICLR 2025 · similarity 0.87)
- LLM Augmented LLMs: Expanding Capabilities through Composition (ICLR 2024 · similarity 0.87)
- Self-Adapting Language Models (NeurIPS 2025 · similarity 0.86)
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.
