Chain-of-Thought as a Lens: Evaluating Structured Reasoning Alignment between Human Preferences and Large Language Models
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866910151237173248 |
|---|---|
| author | Wang, Boxuan Li, Zhuoyun Huang, Xinmiao Huang, Xiaowei Dong, Yi |
| author_facet | Wang, Boxuan Li, Zhuoyun Huang, Xinmiao Huang, Xiaowei Dong, Yi |
| contents | This paper primarily demonstrates a method to quantitatively assess the alignment between multi-step, structured reasoning in large language models and human preferences. We introduce the Alignment Score, a semantic-level metric that compares a model-produced chain of thought traces with a human-preferred reference by constructing semantic-entropy-based matrices over intermediate steps and measuring their divergence. Our analysis shows that Alignment Score tracks task accuracy across models and hop depths, and peaks at 2-hop reasoning. Empirical results further indicate that misalignment at greater reasoning depths is driven mainly by alignment errors such as thematic shift and redundant reasoning. Viewing chain sampling as drawing from a distribution over reasoning paths, we empirically demonstrate a strong and consistent correlation between Alignment Score and accuracy, readability, and coherence, supporting its use as a diagnostic signal. The code is available. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_06168 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Chain-of-Thought as a Lens: Evaluating Structured Reasoning Alignment between Human Preferences and Large Language Models Wang, Boxuan Li, Zhuoyun Huang, Xinmiao Huang, Xiaowei Dong, Yi Artificial Intelligence This paper primarily demonstrates a method to quantitatively assess the alignment between multi-step, structured reasoning in large language models and human preferences. We introduce the Alignment Score, a semantic-level metric that compares a model-produced chain of thought traces with a human-preferred reference by constructing semantic-entropy-based matrices over intermediate steps and measuring their divergence. Our analysis shows that Alignment Score tracks task accuracy across models and hop depths, and peaks at 2-hop reasoning. Empirical results further indicate that misalignment at greater reasoning depths is driven mainly by alignment errors such as thematic shift and redundant reasoning. Viewing chain sampling as drawing from a distribution over reasoning paths, we empirically demonstrate a strong and consistent correlation between Alignment Score and accuracy, readability, and coherence, supporting its use as a diagnostic signal. The code is available. |
| title | Chain-of-Thought as a Lens: Evaluating Structured Reasoning Alignment between Human Preferences and Large Language Models |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2511.06168 |