SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2402.05808 · 2024

Training Large Language Models for Reasoning through Reverse Curriculum Reinforcement Learning

arXiv · PDF · Open in the Atlas

Code that ran

15 of 19 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; 7 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 7 failures of our instrument; 4 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/2402.05808.json · badge: /api/ran/2402.05808.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
woooodyy/llm-reverse-curriculum-rl canonical (classifier label)
classifier basis: census
15 ran · 4 no recorded run
FunctionStatusWhere it lives
calculate_flops Ran
ran, honoured the contract we drafted
woooodyy/llm-reverse-curriculum-rl/R3_others/dschat/utils/perf.py
pointer only (licence: NONE) · get_code("069f56cb00324e10")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

discount_cumsum Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
woooodyy/llm-reverse-curriculum-rl/R3_math/src/utils.py
pointer only (licence: NONE) · get_code("7ffa253a733f9bfc")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

floatify Ran
ran, honoured the contract we drafted
woooodyy/llm-reverse-curriculum-rl/R3_math/src/python_stdout_engine.py
pointer only (licence: NONE) · get_code("f8cd262700389f42")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

format_code Ran
ran; our contract draft was wrong, not the code
woooodyy/llm-reverse-curriculum-rl/R3_math/src/python_stdout_engine.py
pointer only (licence: NONE) · get_code("b425aefc590684dc")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

gather_log_probs Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
woooodyy/llm-reverse-curriculum-rl/R3_others/dschat/rlhf/ppo_trainer.py
pointer only (licence: NONE); the stored copy get_code serves, from zyushun/adam-mini, is pointer only (licence: NONE) · get_code("2f0a690c70d11c23")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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_eval_ds_config Ran
ran; our contract draft was wrong, not the code
woooodyy/llm-reverse-curriculum-rl/R3_others/dschat/utils/ds_utils.py
pointer only (licence: NONE) · get_code("792e87a1badc5642")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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_hf_configs Ran
ran, honoured the contract we drafted
woooodyy/llm-reverse-curriculum-rl/R3_others/dschat/utils/perf.py
pointer only (licence: NONE) · get_code("df20c2f4083d17af")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

is_numeric Ran
ran, violated the contract we drafted
woooodyy/llm-reverse-curriculum-rl/R3_math/src/utils.py
pointer only (licence: NONE); the stored copy get_code serves, from WNQzhu/CARFT, is pointer only (licence: NONE) · get_code("c3b86a3bceec86cd")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

log_init Ran
ran, honoured the contract we drafted
woooodyy/llm-reverse-curriculum-rl/R3_others/dschat/rlhf/rlhf_engine.py
pointer only (licence: NONE) · get_code("429d42a3bf925da5")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

number_it Ran
ran, honoured the contract we drafted
woooodyy/llm-reverse-curriculum-rl/R3_math/src/python_stdout_engine.py
pointer only (licence: NONE) · get_code("1af181bffd7c0355")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

only_optimize_lora_parameters Ran
ran; our contract draft was wrong, not the code
woooodyy/llm-reverse-curriculum-rl/R3_others/dschat/utils/module/lora.py
pointer only (licence: NONE) · get_code("7d082cd1b7edc2aa")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

process_code Ran
ran; our contract draft was wrong, not the code
woooodyy/llm-reverse-curriculum-rl/R3_math/src/python_engine.py
pointer only (licence: NONE) · get_code("2221a9c4d6aaccc3")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

run_python_code Ran
ran, honoured the contract we drafted
woooodyy/llm-reverse-curriculum-rl/R3_math/src/python_engine.py
pointer only (licence: NONE) · get_code("23907c9b2306bf66")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

test_answer_for_reward Ran
ran, honoured the contract we drafted
woooodyy/llm-reverse-curriculum-rl/R3_others/dschat/rlhf/ppo_trainer.py
pointer only (licence: NONE) · get_code("add7bb376d093e6a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

to_device Ran
ran; our contract draft was wrong, not the code
woooodyy/llm-reverse-curriculum-rl/R3_others/dschat/utils/utils.py
pointer only (licence: NONE) · get_code("387b9e6c4f21dd62")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

causal_lm_model_to_fp32_loss Not yet run
unverified (harvested, no recorded run)
woooodyy/llm-reverse-curriculum-rl/R3_others/dschat/utils/model/model_utils.py
pointer only (licence: NONE); the stored copy get_code serves, from zmzhang2000/trustworthy-alignment, is pointer only (licence: NONE) · get_code("b6934717c02ba140")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

do_gather Not yet run
unverified (harvested, no recorded run)
woooodyy/llm-reverse-curriculum-rl/R3_math/src/utils.py
pointer only (licence: NONE); the stored copy get_code serves, from WNQzhu/CARFT, is pointer only (licence: NONE) · get_code("32851be78f5ecd4c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

get_tokenizer Not yet run
unverified (harvested, no recorded run)
woooodyy/llm-reverse-curriculum-rl/R3_others/dschat/utils/utils.py
pointer only (licence: NONE) · get_code("60da16ae2ccbd130")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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.

load_hf_tokenizer Not yet run
unverified (harvested, no recorded run)
woooodyy/llm-reverse-curriculum-rl/R3_others/dschat/utils/utils.py
pointer only (licence: NONE) · get_code("51909067f349309c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, woooodyy/llm-reverse-curriculum-rl 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 19 samples lifted from this paper's linked repositories, 19 carry a function identity and 0 do not. Of the 19 distinct functions with an identity, 12 are also present in other papers whose code we hold (8 with at least one paper dated earlier, 1 the same month, 11 later) and 7 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
gather_log_probs 9 other papers
3 dated earlier, 5 later, 1 same year, order not determinable
2308.02223 (2023-08)
ESRL: Efficient Sampling-based Reinforcement Learning for Sequence Generation
2310.10505 (2023-10)
ReMax: A Simple, Effective, and Efficient Reinforcement Learning Method for Aligning Large Language Models
2024.naacl-long.326 (2024 (year only))
LSTDial: Enhancing Dialogue Generation via Long-and Short-Term Measurement Feedback
and 6 more not named here
log_init 9 other papers
3 dated earlier, 5 later, 1 same year, order not determinable
2308.02223 (2023-08)
ESRL: Efficient Sampling-based Reinforcement Learning for Sequence Generation
2310.10505 (2023-10)
ReMax: A Simple, Effective, and Efficient Reinforcement Learning Method for Aligning Large Language Models
2024.naacl-long.326 (2024 (year only))
LSTDial: Enhancing Dialogue Generation via Long-and Short-Term Measurement Feedback
and 6 more not named here
calculate_flops 6 other papers
1 dated earlier, 5 later
2310.10505 (2023-10)
ReMax: A Simple, Effective, and Efficient Reinforcement Learning Method for Aligning Large Language Models
2403.14112 (2024-03)
Benchmarking Chinese Commonsense Reasoning of LLMs: From Chinese-Specifics to Reasoning-Memorization Correlations
2405.16455 (2024-05)
On the Algorithmic Bias of Aligning Large Language Models with RLHF: Preference Collapse and Matching Regularization
and 3 more not named here
get_hf_configs 6 other papers
1 dated earlier, 5 later
2310.10505 (2023-10)
ReMax: A Simple, Effective, and Efficient Reinforcement Learning Method for Aligning Large Language Models
2403.14112 (2024-03)
Benchmarking Chinese Commonsense Reasoning of LLMs: From Chinese-Specifics to Reasoning-Memorization Correlations
2405.16455 (2024-05)
On the Algorithmic Bias of Aligning Large Language Models with RLHF: Preference Collapse and Matching Regularization
and 3 more not named here
to_device 6 other papers
1 dated earlier, 1 the same month, 3 later, 1 same year, order not determinable
2310.20246 (2023-10)
Breaking Language Barriers in Multilingual Mathematical Reasoning: Insights and Observations
2024.naacl-long.326 (2024 (year only))
LSTDial: Enhancing Dialogue Generation via Long-and Short-Term Measurement Feedback
2402.18150 (2024-02)
Unsupervised Information Refinement Training of Large Language Models for Retrieval-Augmented Generation
and 3 more not named here
do_gather 3 other papers
1 dated earlier, 2 later
2401.08967 (2024-01)
REFT: Reasoning with REinforced Fine-Tuning
2404.03592 (2024-04)
ReFT: Representation Finetuning for Language Models
2508.15868 (2025-08)
CARFT: Boosting LLM Reasoning via Contrastive Learning with Annotated Chain-of-Thought-based Reinforced Fine-Tuning
floatify 3 other papers
3 later
2412.01981 (2024-12)
Free Process Rewards without Process Labels
2025.findings-acl.605 (2025 (year only))
LEMMA: Learning from Errors for MatheMatical Advancement in LLMs
2505.21413 (2025-05)
RefTool: Reference-Guided Tool Creation for Knowledge-Intensive Reasoning
get_eval_ds_config 3 other papers
3 later
2405.16455 (2024-05)
On the Algorithmic Bias of Aligning Large Language Models with RLHF: Preference Collapse and Matching Regularization
2410.16843 (2024-10)
Trustworthy Alignment of Retrieval-Augmented Large Language Models via Reinforcement Learning
2410.21662 (2024-10)
$f$-PO: Generalizing Preference Optimization with $f$-divergence Minimization
is_numeric 3 other papers
1 dated earlier, 2 later
2401.08967 (2024-01)
REFT: Reasoning with REinforced Fine-Tuning
2404.03592 (2024-04)
ReFT: Representation Finetuning for Language Models
2508.15868 (2025-08)
CARFT: Boosting LLM Reasoning via Contrastive Learning with Annotated Chain-of-Thought-based Reinforced Fine-Tuning
causal_lm_model_to_fp32_loss 2 other papers
2 later
2405.16455 (2024-05)
On the Algorithmic Bias of Aligning Large Language Models with RLHF: Preference Collapse and Matching Regularization
2410.16843 (2024-10)
Trustworthy Alignment of Retrieval-Augmented Large Language Models via Reinforcement Learning
only_optimize_lora_parameters 2 other papers
2 later
2405.16455 (2024-05)
On the Algorithmic Bias of Aligning Large Language Models with RLHF: Preference Collapse and Matching Regularization
2410.21662 (2024-10)
$f$-PO: Generalizing Preference Optimization with $f$-divergence Minimization
format_code 1 other paper
1 dated earlier
2401.06081 (2024-01)
Improving Large Language Models via Fine-grained Reinforcement Learning with Minimum Editing Constraint

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

In this paper, we propose R$^3$: Learning Reasoning through Reverse Curriculum Reinforcement Learning (RL), a novel method that employs only outcome supervision to achieve the benefits of process supervision for large language models. The core challenge in applying RL to complex reasoning is to identify a sequence of actions that result in positive rewards and provide appropriate supervision for optimization. Outcome supervision provides sparse rewards for final results without identifying error locations, whereas process supervision offers step-wise rewards but requires extensive manual annotation. R$^3$ overcomes these limitations by learning from correct demonstrations. Specifically, R$^3$ progressively slides the start state of reasoning from a demonstration's end to its beginning, facilitating easier model exploration at all stages. Thus, R$^3$ establishes a step-wise curriculum, allowing outcome supervision to offer step-level signals and precisely pinpoint errors. Using Llama2-7B, our method surpasses RL baseline on eight reasoning tasks by $4.1$ points on average. Notebaly, in program-based reasoning on GSM8K, it exceeds the baseline by $4.2$ points across three backbone models, and without any extra data, Codellama-7B + R$^3$ performs comparable to larger models or closed-source models.

For agents

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

get_harvested_code_for_paper("2402.05808")
get_code_for_paper("2402.05808")
have("2402.05808")

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

curl https://syntology.ai/api/ran/2402.05808.json

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

[![Syntology run record](https://syntology.ai/api/ran/2402.05808.svg)](https://syntology.ai/paper/2402.05808)

Connect an agent — have() is free.