Papers › Direct Preference Optimization: Your Language Model is Secretly a Reward Model

Direct Preference Optimization: Your Language Model is Secretly a Reward Model

29 May 2023NeurIPS 2023 11arXiv:2305.18290archive 2025-07-28

Rafael Rafailov, Archit Sharma, Eric Mitchell, Stefano Ermon, Christopher D. Manning, Chelsea Finn

While large-scale unsupervised language models (LMs) learn broad world knowledge and some reasoning skills, achieving precise control of their behavior is difficult due to the completely unsupervised nature of their training. Existing methods for gaining such steerability collect human labels of the relative quality of model generations and fine-tune the unsupervised LM to align with these preferences, often with reinforcement learning from human feedback (RLHF). However, RLHF is a complex and often unstable procedure, first fitting a reward model that reflects the human preferences, and then fine-tuning the large unsupervised LM using reinforcement learning to maximize this estimated reward without drifting too far from the original model. In this paper we introduce a new parameterization of the reward model in RLHF that enables extraction of the corresponding optimal policy in closed form, allowing us to solve the standard RLHF problem with only a simple classification loss. The resulting algorithm, which we call Direct Preference Optimization (DPO), is stable, performant, and computationally lightweight, eliminating the need for sampling from the LM during fine-tuning or performing significant hyperparameter tuning. Our experiments show that DPO can fine-tune LMs to align with human preferences as well as or better than existing methods. Notably, fine-tuning with DPO exceeds PPO-based RLHF in ability to control sentiment of generations, and matches or improves response quality in summarization and single-turn dialogue while being substantially simpler to implement and train.

PaperPDFConference PDFCodeCode Syntology ranMethod as code

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 31 code samples harvested from 7 repositories linked to this paper: 25 ran (of which 1 constructed an object rather than computing a result; 21 with no instrument failure: 0 honoured, 0 violated, 21 with no contract checked; 4 where Syntology's instrument failed) · 6 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: community (archive-listed): 31 samples from 7 repositories, 25 ran (21 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.

No sample is linked to this paper from 15 of its repositories that Syntology harvested for other papers:

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.

29 repositories listed; official and paper-mentioned ones first.

bitbucket.org/paladinanalytics/direct-preference-optimization mentioned on GitHubpytorchSyntology: not harvested report
12kimih/hicupid mentioned on GitHubpytorchApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Exploring the Potential of LLMs as Personalized Assistants: Dataset, Evaluation, and Analysis (arXiv:2506.01262; that paper's own run record: 6 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
KomeijiForce/Active_Passive_Constraint_Koishiday_2024 mentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Quantifying and Optimizing Global Faithfulness in Persona-driven Role-playing (arXiv:2405.07726; that paper's own run record: 4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified); nothing checks that they implement this paper's method. report
TU2021/DPO-VP mentioned on GitHubpytorchSyntology: not harvested report
ahmed-alllam/Direct-Preference-Optimization mentioned on GitHubpytorchSyntology: not harvested report
babelscape/alert mentioned on GitHubpytorchNOASSERTIONSyntology: not harvested report
davidkim205/komt mentioned on GitHubpytorchApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Self-Alignment with Instruction Backtranslation (arXiv:2308.06259; that paper's own run record: 7 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 1 where Syntology's instrument failed) · 1 unverified); nothing checks that they implement this paper's method. report
eric-mitchell/direct-preference-optimization mentioned on GitHubpytorchApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to On Extending Direct Preference Optimization to Accommodate Ties (arXiv:2409.17431; that paper's own run record: 5 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 4 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
frankxwang/dpo-prefix-sharing mentioned on GitHubpytorchSyntology: not harvested report
huggingface/trl mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Filtered Direct Preference Optimization (arXiv:2404.13846; that paper's own run record: 2 ran (of which 2 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified; every one of the 2 samples that ran constructed an object rather than computing a result); nothing checks that they implement this paper's method. report
icip-cas/Word-Deed-Consistency-Test mentioned on GitHubpytorchSyntology: 1 sample, 1 ran (1 instrument-failed) · 0 unverified report
imagination-research/lbt mentioned on GitHubMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Can LLMs Learn by Teaching for Better Reasoning? A Preliminary Study (arXiv:2406.14629; that paper's own run record: 3 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 1 violated, 1 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
internlm/xtuner mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to LoRA: Low-Rank Adaptation of Large Language Models (arXiv:2106.09685; that paper's own run record: 51 ran (of which 19 constructed an object rather than computing a result; 43 with no instrument failure: 1 honoured, 0 violated, 42 with no contract checked; 8 where Syntology's instrument failed) · 33 unverified); nothing checks that they implement this paper's method. report
kidist-amde/ddro mentioned on GitHubpytorchSyntology: 2 samples, 2 ran (0 instrument-failed) · 0 unverified report
kidist-amde/ddro-direct-document-relevance-optimization mentioned on GitHubpytorchApache-2.0Syntology: 12 samples, 12 ran (0 instrument-failed) · 0 unverified report
lucidrains/self-rewarding-lm-pytorch mentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Self-Rewarding Language Models (arXiv:2401.10020; that paper's own run record: 6 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 2 violated, 3 with no contract checked; 1 where Syntology's instrument failed) · 5 unverified); nothing checks that they implement this paper's method. report
lujunru/sampo mentioned on GitHubpytorchMITSyntology: not harvested report
lvwerra/trl mentioned on GitHubpytorchApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Intent-conditioned and Non-toxic Counterspeech Generation using Multi-Task Instruction... (arXiv:2403.10088; that paper's own run record: 9 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 4 violated, 2 with no contract checked; 3 where Syntology's instrument failed) · 1 unverified); nothing checks that they implement this paper's method. report
padlex/trl mentioned on GitHubpytorchApache-2.0Syntology: 12 samples, 7 ran (3 instrument-failed) · 5 unverified report
sauc-abadal/alt mentioned on GitHubpytorchSyntology: 2 samples, 1 ran (0 instrument-failed) · 1 unverified report
shibing624/medicalgpt mentioned on GitHubpytorchApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to ORPO: Monolithic Preference Optimization without Reference Model (arXiv:2403.07691; that paper's own run record: 11 ran (of which 1 constructed an object rather than computing a result; 9 with no instrument failure: 1 honoured, 0 violated, 8 with no contract checked; 2 where Syntology's instrument failed) · 3 unverified); nothing checks that they implement this paper's method. report
sunzeyeah/rlhf mentioned on GitHubpytorchSyntology: 1 sample, 1 ran (0 instrument-failed) · 0 unverified report
tatsu-lab/alpaca_farm mentioned on GitHubpytorchApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Beyond Reward Hacking: Causal Rewards for Large Language Model Alignment (arXiv:2501.09620; that paper's own run record: 11 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 3 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method. report
uppaal/detox-edit mentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Model Editing as a Robust and Denoised variant of DPO: A Case Study on Toxicity (arXiv:2405.13967; that paper's own run record: 7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 5 unverified); nothing checks that they implement this paper's method. report
wangclnlp/deepspeed-chat-extension mentioned on GitHubpytorchApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to ESRL: Efficient Sampling-based Reinforcement Learning for Sequence Generation (arXiv:2308.02223; that paper's own run record: 12 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 0 violated, 11 with no contract checked; 1 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method. report
yifan-song793/eto mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Trial and Error: Exploration-Based Trajectory Optimization for LLM Agents (arXiv:2403.02502; that paper's own run record: 5 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 4 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method. report
2024-MindSpore-1/Code3 mindsporeSyntology: not harvested report
OctopusMind/DPO pytorchSyntology: 1 sample, 1 ran (0 instrument-failed) · 0 unverified report
yangyucheng000/Paper-3 mindsporeSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to DPSUR: Accelerating Differentially Private Stochastic Gradient Descent Using Selective... (arXiv:2311.14056; that paper's own run record: 4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 7 unverified); nothing checks that they implement this paper's method. 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.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

31 samples harvested: 25 ran (of which 1 constructed an object rather than computing a result; 21 with no instrument failure: 0 honoured, 0 violated, 21 with no contract checked; 4 where Syntology's instrument failed) · 6 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.

2ran · our draft was wrong
2ran; the contract check could not run
21ran
6unverified

Licence: 2 of the 31 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 7 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “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. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

icip-cas/Word-Deed-Consistency-Test community (archive-listed) 1 sample, 1 ran (1 instrument-failed) · 0 unverified
preference_loss in 9 papers dpo/trainers.py ran; the contract check could not run (RAISES) fingerprinted no licence file found · pointer only · c79211fc0225a27f · report fixture: signature policy_chosen_logps:Tensor/1/float, policy_rejected_logps:Tensor/1/float, reference_chosen_logps:Tensor/1/float, reference_rejected_logps:Tensor/1/float, beta:float/0/float, label_smoothing:float/0/float, ipo:bool/0/bool, reference_free:bool/0/bool · returned tuple · level recorded 2026-09-22
kidist-amde/ddro community (archive-listed) 2 samples, 2 ran (0 instrument-failed) · 0 unverified
T5ForPretrain src/pretrain/T5ForPretrain.py ran Apache-2.0 (permissive) · 88b1660f36f94aca · report fixture: signature input_ids:Tensor/2/int, attention_mask:Tensor/2/float, decoder_input_ids:Tensor/2/int, decoder_attention_mask:Tensor/2/float, labels:Tensor/2/int, return_dict:bool/0/bool · returned Seq2SeqLMOutput · level recorded 2026-09-24
T5ForPretrainDPO src/pretrain/T5ForPretrain.py ran Apache-2.0 (permissive) · 9fc13a517d28b68e · report fixture: signature input_ids:Tensor/2/int, attention_mask:Tensor/2/int, decoder_input_ids:Tensor/2/int, decoder_attention_mask:Tensor/2/int, labels:Tensor/2/int, return_dict:bool/0/bool · returned Seq2SeqLMOutput · level recorded 2026-09-24
kidist-amde/ddro-direct-document-relevance-optimization community (archive-listed) 12 samples, 12 ran (0 instrument-failed) · 0 unverified
average_precision in 38 papers src/utils/evaluate.py ran Apache-2.0 (permissive) · 99d26fdcc1af5b62 · report fixture: signature truth:set/0/?, pred:list/1/str · returned float · level recorded 2026-09-25
compare_metrics src/utils/compare_models.py ran Apache-2.0 (permissive) · 9169d5b42acf842a · report fixture: signature metrics_df_1:DataFrame/0/?, metrics_df_2:DataFrame/0/?, metrics_list:list/1/str, test:str/0/str · returned DataFrame · level recorded 2026-09-24
hit_at_k src/utils/evaluate.py ran Apache-2.0 (permissive) · e9ac5f8dce946d2f · report fixture: signature truth:list/1/int, pred:list/1/int, k:int/0/int · returned float · level recorded 2026-09-24
load_checkpoint in 277 papers src/pretrain/train_ddro_encoder_decoder.py ran Apache-2.0 (permissive) · 0f3aeaf1bff007f2 · report fixture: signature path:str/0/str · returned dict · level recorded 2026-09-24
load_data in 700 papers src/pretrain/eval_ddro_docid_ranking.py ran Apache-2.0 (permissive) · 4099bb8045c005b6 · report fixture: signature file_path:str/0/str · returned list · level recorded 2026-09-24
load_doc_offsets src/pretrain/train_ddro_encoder_decoder.py ran Apache-2.0 (permissive) · 01a7e0dc6f29f042 · report fixture: signature path:str/0/str · returned dict · level recorded 2026-09-24
load_encoded_docid src/utils/generate_demo.py ran Apache-2.0 (permissive) · 9c879833332cded9 · report fixture: signature docid_path:str/0/str · returned tuple · level recorded 2026-09-24
load_metrics in 6 papers src/utils/compare_models.py ran Apache-2.0 (permissive) · 1233495c1bf99438 · report fixture: signature file_path:str/0/str · returned DataFrame · level recorded 2026-09-24
load_query_texts src/pretrain/train_ddro_encoder_decoder.py ran Apache-2.0 (permissive) · 8c68bf8c060f4814 · report fixture: signature path:str/0/str · returned dict · level recorded 2026-09-24
prepare_dataset in 32 papers src/pretrain/finetune_docTTTTTquery_NQ.py ran Apache-2.0 (permissive) · 19b1210f13924026 · report fixture: signature data_path:str/0/str, tokenizer:callable/0/?, max_length:int/0/int · returned Dataset · level recorded 2026-09-24
recall_at_k in 31 papers src/utils/evaluate.py ran Apache-2.0 (permissive) · a05d09685378cdb9 · report fixture: signature truth:set/0/?, pred:list/1/int, k:int/0/int · returned float · level recorded 2026-09-24
split_dataset in 65 papers src/pretrain/finetune_docTTTTTquery_NQ.py ran Apache-2.0 (permissive) · bdad0f787751d773 · report fixture: signature dataset:Dataset/0/?, test_size:float/0/float · returned tuple · level recorded 2026-09-24
padlex/trl community (archive-listed) 12 samples, 7 ran (3 instrument-failed) · 5 unverified
conversations_formatting_function trl/extras/dataset_formatting.py ran Apache-2.0 (permissive) · e491afec66bfd582 · report fixture: signature tokenizer:callable/0/?, messages_field:str/0/str · returned function · level recorded 2026-09-24
convert_state_dict in 13 papers trl/models/sd_utils.py ran · our draft was wrong Apache-2.0 (permissive) · 8f1ce89b816a0734 · report fixture: signature state_dict:dict/0/?, mapping:dict/0/? · returned dict · level recorded 2026-09-24
convert_to_scalar in 24 papers trl/core.py ran · our draft was wrong Apache-2.0 (permissive) · c32b636f13ed9ee5 · report fixture: signature stats:dict/0/? · returned dict · level recorded 2026-09-25
get_all_parameters in 7 papers trl/models/utils.py ran Apache-2.0 (permissive) · 2f5119732e8dbd63 · report fixture: signature sub_module:callable/0/?, recurse:bool/0/bool · returned chain · level recorded 2026-09-25
instructions_formatting_function trl/extras/dataset_formatting.py ran Apache-2.0 (permissive) · 875b816ac9842673 · report fixture: signature tokenizer:MockTokenizer/0/? · returned function · level recorded 2026-09-24
iter_params trl/models/utils.py ran Apache-2.0 (permissive) · 48685a350888e505 · report fixture: signature module:callable/0/?, recurse:bool/0/bool · returned list · level recorded 2026-09-25
top_k_top_p_filtering in 96 papers
census: 8 implementations from 8 papers, 4 distinct outputs across 2 buckets, one shared input per bucket
trl/core.py ran; the contract check could not run (RAISES) Apache-2.0 (permissive) · 5bcf8e086423cf6b · report fixture: signature logits:Tensor/2/float, top_k:int/0/int, top_p:float/0/float, filter_value:float/0/float, min_tokens_to_keep:int/0/int · returned Tensor · level recorded 2026-09-24
convert_state_dict_to_diffusers trl/models/sd_utils.py unverified Apache-2.0 (permissive) · fa59c920caf016e6 · report
flatten_dict in 72 papers trl/core.py unverified Apache-2.0 (permissive) · a51ae45a1f6c0b5b · report
is_transformers_greater_than in 9 papers trl/import_utils.py unverified Apache-2.0 (permissive) · 2a9ca02cb6e196ff · report
scheduler_step in 5 papers trl/models/modeling_sd_base.py unverified Apache-2.0 (permissive) · ade247252fd2d3db · report
setup_chat_format trl/models/utils.py unverified Apache-2.0 (permissive) · ccaedb0c5f904b0b · report
sauc-abadal/alt community (archive-listed) 2 samples, 1 ran (0 instrument-failed) · 1 unverified
Policy in 23 papers alt/trainer/alt_trainer.py ran MIT (permissive) · 0a800ad29eb472ac · report fixture: returned Policy · level recorded 2026-09-22
ALTTrainer_noKL alt/trainer/alt_trainer.py unverified MIT (permissive) · 68d8bd0d51b34185 · report
sunzeyeah/rlhf community (archive-listed) 1 sample, 1 ran (0 instrument-failed) · 0 unverified
PairWiseLoss src/models/loss.py ran · metamorphic tier: invariant fingerprinted no licence file found · pointer only · d69bfa3e1e60a92b · report fixture: signature chosen_reward:Tensor/2/float, reject_reward:Tensor/2/float · returned Tensor · level recorded 2026-09-22
OctopusMind/DPO community (archive-listed) 1 sample, 1 ran (0 instrument-failed) · 0 unverified
DPO dpo.py ran Apache-2.0 (permissive) · be7453ec96f94038 · report fixture: signature inputs_ids:Tensor/2/int, attention_mask:Tensor/2/float, ref_logits:Tensor/3/float, labels_mask:Tensor/2/int · returned NoneType · level recorded 2026-09-24

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it. Siblings are assemblies lifted from this paper in the same batch for the same method: alternative readings kept side by side; none replaces or supersedes another.

5 assemblies on this paper: 5 harvested, 0 generated, listed in 1 groups (siblings lifted in one batch form one group).

5 sibling assemblies for Direct Preference Optimization (DPO) (harvested), lifted in the same batch on 2026-09-09: alternative readings kept side by side; none replaces or supersedes another.

harvested Assembly for Direct Preference Optimization (DPO) (graph method: Direct Preference Optimization (DPO)). 2 bricks; individually on a fixture: 2 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. The assembly has not been run end-to-end.

one of 5 siblings

  1. T5ForPretrainDPO entry · level 1 · ran · 9fc13a517d28b68e
  2. T5ForPretrain helper · level 1 · ran · 88b1660f36f94aca

harvested Assembly for Direct Preference Optimization (DPO) (graph method: Direct Preference Optimization (DPO)). 6 bricks; individually on a fixture: 5 ran (of which 1 constructed an object rather than computing a result; 3 with no instrument failure: 2 honoured, 0 violated, 1 with no contract checked; 2 where Syntology's instrument failed) · 1 unverified. The assembly has not been run end-to-end.

one of 5 siblings

  1. ALTTrainer_noKL entry · level 0 · unverified · 68d8bd0d51b34185
  2. Policy model · level 1 · ran · 0a800ad29eb472ac
  3. reduce_mean helper · level 1 · ran · our draft was wrong · e5bacba887a721ef
  4. logits_to_entropy helper · level 1 · ran · honoured contract · 601afc113e2c5295
  5. mask_pad helper · level 1 · ran · honoured contract · a9c6a287d240022c
  6. reduce_sum helper · level 1 · ran · our draft was wrong · 5874b0a8910897fd

harvested Assembly for Direct Preference Optimization (DPO) (graph method: Direct Preference Optimization (DPO)). 1 brick; individually on a fixture: 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. The assembly has not been run end-to-end.

one of 5 siblings

  1. DPO entry · level 1 · ran · be7453ec96f94038

harvested Assembly for Direct Preference Optimization (DPO) (graph method: Direct Preference Optimization (DPO)). 1 brick; individually on a fixture: 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. The assembly has not been run end-to-end.

one of 5 siblings

  1. PairWiseLoss entry · level 1 · ran · d69bfa3e1e60a92b

harvested Assembly for Direct Preference Optimization (DPO) (graph method: Direct Preference Optimization (DPO)). 1 brick; individually on a fixture: 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified. The assembly has not been run end-to-end.

one of 5 siblings

  1. preference_loss entry · level 1 · ran; the contract check could not run (RAISES) · c79211fc0225a27f

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2305.18290") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Language ModelingLanguage ModellingPreference MappingWorld Knowledgemodelreinforcement-learning

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Introduced by this paper: DPO

ALIGNDPO

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