Papers › Unveiling the Mechanisms of Explicit CoT Training: How CoT Enhances Reasoning Generalization

Unveiling the Mechanisms of Explicit CoT Training: How CoT Enhances Reasoning Generalization

7 Feb 2025arXiv:2502.04667archive 2025-07-28

Xinhao Yao, Ruifeng Ren, Yun Liao, Yong liu

The integration of explicit Chain-of-Thought (CoT) reasoning into training large language models (LLMs) has advanced their reasoning capabilities, yet the mechanisms by which CoT enhances generalization remain poorly understood. This work investigates (1) how CoT training reshapes internal model representations and (2) why it improves both in-distribution (ID) and out-of-distribution (OOD) reasoning generalization. Through controlled experiments and theoretical analysis, we derive the following key insights. 1) Structural Advantage: CoT training internalizes reasoning into a two-stage generalizing circuit, where the number of stages corresponds to the explicit reasoning steps during training. Notably, CoT-trained models resolve intermediate results at shallower layers compared to non-CoT counterparts, freeing up deeper layers to specialize in subsequent reasoning steps. 2) Theoretical Analysis: the information-theoretic generalization bounds via distributional divergence can be decomposed into ID and OOD components. While ID error diminishes with sufficient training regardless of CoT, OOD error critically depends on CoT: Non-CoT training fails to generalize to OOD samples due to unseen reasoning patterns, whereas CoT training achieves near-perfect OOD generalization by mastering subtasks and reasoning compositions during training. The identified mechanisms explain our experimental results: CoT training accelerates convergence and enhances generalization from ID to both ID and OOD scenarios while maintaining robust performance even with tolerable noise. These findings are further validated on complex real-world datasets. This paper offers valuable insights for designing CoT strategies to enhance LLM reasoning robustness.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 10 code samples harvested from 1 repository linked to this paper: 9 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 1 honoured, 0 violated, 4 with no contract checked; 4 where Syntology's instrument failed) · 1 unverified. Where Syntology's instrument failed: 2 where our contract draft was wrong; 2 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 10 samples from 1 repository, 9 ran (5 with no instrument failure; 4 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-22 and 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

chen123ctrls/t-cotmechanism officialmentioned in paperjaxMITSyntology: 10 samples, 9 ran (4 instrument-failed) · 1 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

10 samples harvested: 9 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 1 honoured, 0 violated, 4 with no contract checked; 4 where Syntology's instrument failed) · 1 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · honoured contract
2ran · our draft was wrong
2ran; the contract check could not run
4ran
1unverified

Licence: 4 of the 10 samples are pointer only, meaning Syntology does not serve that copy's text. For 4 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from chen123ctrls/t-cotmechanism. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

chen123ctrls/t-cotmechanism official repository 10 samples, 9 ran (4 instrument-failed) · 1 unverified
add_step_noise_to_gsm_fobar RealisticDataVerification/data_process.py ran MIT (permissive) · cabf394868b62573 · report fixture: signature data:dict/0/?, error_prob:float/0/float · returned dict · level recorded 2026-09-24
batch_data in 20 papers RealisticDataVerification/utils/gen_vllm.py ran · our draft was wrong fingerprinted MIT (permissive) here; the stored copy is pointer only · d49412b3a690206e · report fixture: signature data_list:list/1/Tensor, batch_size:int/0/int · returned list · level recorded 2026-09-25
eval_items ControlledExperiments/eval_qa.py ran MIT (permissive) · a3ae250c1e40eeef · report fixture: signature all_items:list/1/dict, partition_atomic:bool/0/bool, test_entities:NoneType/0/? · returned dict · level recorded 2026-09-24
eval_res ControlledExperiments/eval_qa.py ran · honoured contract fingerprinted MIT (permissive) · bedd2bb861320f0e · report fixture: signature a:str/0/str, b:str/0/str · returned int · level recorded 2026-09-22
get_config_class_from_processor_class in 29 papers ControlledExperiments/transformers/utils/create_dummy_models.py ran; the contract check could not run (RAISES) MIT (permissive) here; the stored copy is pointer only · 3d999b95009c7537 · report fixture: signature processor_class:callable/0/? · returned type · level recorded 2026-09-25
get_processor_types_from_config_class in 29 papers ControlledExperiments/transformers/utils/create_dummy_models.py ran; the contract check could not run (RAISES) MIT (permissive) here; the stored copy is pointer only · 0c0b81b081919d8e · report fixture: signature config_class:callable/0/?, allowed_mappings:list/1/str · returned tuple · level recorded 2026-09-25
modify_output RealisticDataVerification/data_process.py ran MIT (permissive) · bc440fba1f01ba1e · report fixture: signature example:dict/0/? · returned dict · level recorded 2026-09-25
parse_target ControlledExperiments/eval_qa.py ran · our draft was wrong MIT (permissive) · 62ef5378e10da132 · report fixture: signature target_text:str/0/str · returned tuple · level recorded 2026-09-22
read_data_source_target ControlledExperiments/utils.py ran MIT (permissive) · aea562a62336c08f · report fixture: signature file_name:str/0/str, return_num:bool/0/bool, return_json:bool/0/bool · returned DataFrame · level recorded 2026-09-24
get_architectures_from_config_class in 29 papers ControlledExperiments/transformers/utils/create_dummy_models.py unverified MIT (permissive) here; the stored copy is pointer only · b176efb26f72e507 · report

Tasks

Generalization BoundsSystematic Generalization

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections