Jiaqi Wang, Wanxiang Che, Qiguang Chen, Libo Qin, Jinxuan Zhou
We lifted 15 functions out of this paper's own repositories and ran 12 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| LightChen233/reasoning-boundary | canonical | 7 of 8 |
| lightchen233/reasoning-boundary | canonical | 5 of 7 |
| Function | Status | Where it lives |
|---|---|---|
| RequestOutput | Ran | lightchen233/reasoning-boundary/draw_bound_text.py pointer only (licence: NONE) · get_code("65d7d13bb3815dc4") |
| create_prompt | Ran | LightChen233/reasoning-boundary/request_text.py pointer only (licence: NONE) · get_code("46be0dbb921f8d85") |
| encode_image | Ran | LightChen233/reasoning-boundary/request_multimodal.py pointer only (licence: NONE) · get_code("f41cb1a19b154297") |
| evaluate_expression | Ran | lightchen233/reasoning-boundary/draw_bound_text.py pointer only (licence: NONE) · get_code("82a44b30698a9d51") |
| get_combined_boundary | Ran | lightchen233/reasoning-boundary/draw_bound_text.py pointer only (licence: NONE) · get_code("0d6f01223a82f7bb") |
| get_combined_boundary | Ran | LightChen233/reasoning-boundary/utils/tools.py pointer only (licence: NONE) · get_code("fc9680a75b75c9f0") |
| judge_error | Ran | lightchen233/reasoning-boundary/draw_bound_text.py pointer only (licence: NONE) · get_code("046fea806c599bb1") |
| loop_judge | Ran | LightChen233/reasoning-boundary/evaluate_multimodal.py pointer only (licence: NONE) · get_code("13bbf36cd13ac257") |
| read_jsonl | Ran | LightChen233/reasoning-boundary/utils/mm_tool.py pointer only (licence: NONE) · get_code("8f0ae0bd1b3d8a66") |
| read_jsonl | Ran | lightchen233/reasoning-boundary/draw_bound_text.py pointer only (licence: NONE) · get_code("86262555b2a8871d") |
| sort_dict | Ran | LightChen233/reasoning-boundary/utils/mm_tool.py pointer only (licence: NONE) · get_code("cf9a53d1a386905c") |
| sort_metric | Ran | LightChen233/reasoning-boundary/utils/mm_tool.py pointer only (licence: NONE) · get_code("2094a3aa5ed68048") |
| create_prompt | Not yet run | LightChen233/reasoning-boundary/request_multimodal.py pointer only (licence: NONE) · get_code("419ea57087ee3539") |
| run | Not yet run | lightchen233/reasoning-boundary/draw_bound_text.py pointer only (licence: NONE) · get_code("3bb6131ac003729b") |
| write_jsonl | Not yet run | lightchen233/reasoning-boundary/draw_bound_text.py pointer only (licence: NONE) · get_code("04733255ce673771") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Chain-of-Thought (CoT) reasoning has emerged as a promising approach for enhancing the performance of large language models (LLMs) on complex reasoning tasks. Recently, a series of studies attempt to explain the mechanisms underlying CoT, aiming to deepen the understanding of its efficacy. Nevertheless, the existing research faces two major challenges: (1) a lack of quantitative metrics to assess CoT capabilities and (2) a dearth of guidance on optimizing CoT performance. Motivated by this, in this work, we introduce a novel reasoning boundary framework (RBF) to address these challenges. To solve the lack of quantification, we first define a reasoning boundary (RB) to quantify the upper-bound of CoT and establish a combination law for RB, enabling a practical quantitative approach applicable to various real-world CoT tasks. To address the lack of optimization, we propose three categories of RBs. We further optimize these categories with combination laws focused on RB promotion and reasoning path optimization for CoT improvement. Through extensive experiments on 27 models and 5 tasks, the study validates the existence and rationality of the proposed framework. Furthermore, it explains the effectiveness of 10 CoT strategies and guides optimization from two perspectives. We hope this work can provide a comprehensive understanding of the boundaries and optimization strategies for reasoning in LLMs. Our code and data are available at https://github.com/LightChen233/reasoning-boundary.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2410.05695")
get_code_for_paper("2410.05695")
have("2410.05695")
Connect an agent — have() is free.