SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2412.04537 · 2024

Understanding Hidden Computations in Chain-of-Thought Reasoning

arXiv · PDF · Open in the Atlas

Code that ran

13 of 14 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 2 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 11 failures of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2412.04537.json · badge: /api/ran/2412.04537.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
rokosbasilisk/filler_tokens canonical (classifier label)
classifier basis: census
13 ran · 1 no recorded run
FunctionStatusWhere it lives
cat_row Ran
ran; our contract draft was wrong, not the code
rokosbasilisk/filler_tokens/src/match3.py
pointer only (licence: NONE) · get_code("5ec0baefcc7f9cec")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

dot_tf_batch_to_type Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
rokosbasilisk/filler_tokens/src/train_match3.py
pointer only (licence: NONE) · get_code("b34a52e418beb7bc")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

find_consecutive_dots_mask Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
rokosbasilisk/filler_tokens/src/train_match3.py
pointer only (licence: NONE) · get_code("6dd6c76d6f6ed809")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_optimizer Ran
ran; our contract draft was wrong, not the code
rokosbasilisk/filler_tokens/src/utils.py
pointer only (licence: NONE) · get_code("1a03d75c0e35e8e7")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

identity_transform Ran
ran; our contract draft was wrong, not the code
rokosbasilisk/filler_tokens/src/match2.py
pointer only (licence: NONE) · get_code("16b09a12d091a303")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

lookup_transform Ran
ran; our contract draft was wrong, not the code
rokosbasilisk/filler_tokens/src/match2.py
pointer only (licence: NONE) · get_code("e8da9e3e5a9a10e5")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

match2_vector_eval_loop Ran
ran; our contract draft was wrong, not the code
rokosbasilisk/filler_tokens/src/train_match2.py
pointer only (licence: NONE) · get_code("84ab83ab5190ada5")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

match2_vector_eval_step Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
rokosbasilisk/filler_tokens/src/train_match2.py
pointer only (licence: NONE) · get_code("441af884cc5dd3fe")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

no_filler_parallel Ran
ran; our contract draft was wrong, not the code
rokosbasilisk/filler_tokens/src/match3.py
pointer only (licence: NONE) · get_code("2c3e6f2e8c98b4c7")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

no_filler_serial Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
rokosbasilisk/filler_tokens/src/match3.py
pointer only (licence: NONE) · get_code("452ad0e734c595f4")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

random_lookup_params Ran
ran, honoured the contract we drafted
rokosbasilisk/filler_tokens/src/match2.py
pointer only (licence: NONE) · get_code("d3cbc2bc4d10c134")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

reshape_crossent Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
rokosbasilisk/filler_tokens/src/utils.py
pointer only (licence: NONE) · get_code("b7c1c5ed38cedeb0")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

train_steps Ran
ran, honoured the contract we drafted
rokosbasilisk/filler_tokens/src/utils.py
pointer only (licence: NONE) · get_code("523caa47bb2d16e6")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

match3_vector_eval_loop Not yet run
unverified (harvested, no recorded run)
rokosbasilisk/filler_tokens/src/train_match3.py
pointer only (licence: NONE) · get_code("0f608d73aca2f8a6")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rokosbasilisk/filler_tokens when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 14 are also present in other papers whose code we hold (14 with at least one paper dated earlier, 0 the same month, 0 later) and 0 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
cat_row 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
dot_tf_batch_to_type 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
find_consecutive_dots_mask 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
get_optimizer 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
identity_transform 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
lookup_transform 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
match2_vector_eval_loop 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
match2_vector_eval_step 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
match3_vector_eval_loop 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
no_filler_parallel 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
no_filler_serial 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
random_lookup_params 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
reshape_crossent 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models
train_steps 1 other paper
1 dated earlier
2404.15758 (2024-04)
Let's Think Dot by Dot: Hidden Computation in Transformer Language Models

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Chain-of-Thought (CoT) prompting has significantly enhanced the reasoning abilities of large language models. However, recent studies have shown that models can still perform complex reasoning tasks even when the CoT is replaced with filler(hidden) characters (e.g., "..."), leaving open questions about how models internally process and represent reasoning steps. In this paper, we investigate methods to decode these hidden characters in transformer models trained with filler CoT sequences. By analyzing layer-wise representations using the logit lens method and examining token rankings, we demonstrate that the hidden characters can be recovered without loss of performance. Our findings provide insights into the internal mechanisms of transformer models and open avenues for improving interpretability and transparency in language model reasoning.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2412.04537")
get_code_for_paper("2412.04537")
have("2412.04537")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2412.04537.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2412.04537.svg)](https://syntology.ai/paper/2412.04537)

Connect an agent — have() is free.