Papers › Beyond Reverse KL: Generalizing Direct Preference Optimization with Diverse Divergence...

Beyond Reverse KL: Generalizing Direct Preference Optimization with Diverse Divergence Constraints

28 Sep 2023arXiv:2309.16240archive 2025-07-28

Chaoqi Wang, Yibo Jiang, Chenghao Yang, Han Liu, Yuxin Chen

The increasing capabilities of large language models (LLMs) raise opportunities for artificial general intelligence but concurrently amplify safety concerns, such as potential misuse of AI systems, necessitating effective AI alignment. Reinforcement Learning from Human Feedback (RLHF) has emerged as a promising pathway towards AI alignment but brings forth challenges due to its complexity and dependence on a separate reward model. Direct Preference Optimization (DPO) has been proposed as an alternative, and it remains equivalent to RLHF under the reverse KL regularization constraint. This paper presents f-DPO, a generalized approach to DPO by incorporating diverse divergence constraints. We show that under certain f-divergences, including Jensen-Shannon divergence, forward KL divergences and α-divergences, the complex relationship between the reward and optimal policy can also be simplified by addressing the Karush-Kuhn-Tucker conditions. This eliminates the need for estimating the normalizing constant in the Bradley-Terry model and enables a tractable mapping between the reward function and the optimal policy. Our approach optimizes LLMs to align with human preferences in a more efficient and supervised manner under a broad set of divergence constraints. Empirically, adopting these divergences ensures a balance between alignment performance and generation diversity. Importantly, f-DPO outperforms PPO-based methods in divergence efficiency, and divergence constraints directly influence expected calibration error (ECE).

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 17 code samples harvested from 1 repository linked to this paper: 16 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; 5 where Syntology's instrument failed) · 1 unverified. Where Syntology's instrument failed: 5 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: community (archive-listed): 17 samples from 1 repository, 16 ran (11 with no instrument failure; 5 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.

alecwangcq/f-divergence-dpo mentioned on GitHubpytorchApache-2.0Syntology: 17 samples, 16 ran (5 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

17 samples harvested: 16 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; 5 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.

5ran · our draft was wrong
11ran
1unverified

Licence: 1 of the 17 samples is pointer only, meaning Syntology does not serve that copy's text. For that sample, 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 alecwangcq/f-divergence-dpo. “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.

alecwangcq/f-divergence-dpo community (archive-listed) 17 samples, 16 ran (5 instrument-failed) · 1 unverified
calculate_entropy in 22 papers generation.py ran fingerprinted Apache-2.0 (permissive) · 194f8dc522933267 · report fixture: signature logits:ndarray/2/float · returned float64 · level recorded 2026-09-24
calculate_score in 15 papers cali/hh_cali_eval.py ran Apache-2.0 (permissive) · f2074220eaa2a95d · report fixture: signature model:callable/0/?, input_prompt_ids:Tensor/2/int, input_ids:Tensor/2/int · returned tuple · level recorded 2026-09-24
calculate_self_bleu generation.py ran fingerprinted Apache-2.0 (permissive) · c38b6fc85679b0b0 · report fixture: signature text:str/0/str · returned float · level recorded 2026-09-24
dpo_loss in 21 papers trainers.py ran Apache-2.0 (permissive) · fc5f3e7e6f97a222 · 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, reference_free:bool/0/bool · returned tuple · level recorded 2026-09-24
extract_anthropic_prompt in 16 papers preference_datasets.py ran · our draft was wrong Apache-2.0 (permissive) here; the stored copy is pointer only · 12f6803880f9b17e · report fixture: signature prompt_and_response:str/0/str · returned str · level recorded 2026-09-22
forward_dpo_loss trainers.py ran Apache-2.0 (permissive) · 6ccfbad96ffcc246 · 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, reference_free:bool/0/bool · returned tuple · level recorded 2026-09-24
get_judge_name mt_bench/compute_agreement.py ran fingerprinted Apache-2.0 (permissive) · d24ec6691ff04146 · report fixture: signature judge:str/0/str · returned str · level recorded 2026-09-25
get_local_dir in 14 papers utils.py ran · our draft was wrong Apache-2.0 (permissive) · 8ab43db002bf984d · report fixture: signature prefixes_to_resolve:list/1/str · returned str · level recorded 2026-09-25
get_local_run_dir in 14 papers utils.py ran · our draft was wrong Apache-2.0 (permissive) · 10ab71f27ffd7de8 · report fixture: signature exp_name:str/0/str, local_dirs:list/1/str · returned str · level recorded 2026-09-24
get_mt_bench_votes_data mt_bench/compute_agreement.py ran Apache-2.0 (permissive) · 7745c0e675406cf2 · report fixture: signature raw_votes:list/1/list · returned list · level recorded 2026-09-25
get_remote_file in 16 papers utils.py ran Apache-2.0 (permissive) · e7ce4c19888c6df6 · report fixture: signature remote_path:str/0/str, local_path:str/0/str · returned str · level recorded 2026-09-25
load_judge_prompts mt_bench/common.py ran · our draft was wrong Apache-2.0 (permissive) · 41fd9954784a1330 · report fixture: signature prompt_file:str/0/str · returned dict · level recorded 2026-09-24
load_model_answers in 5 papers mt_bench/common.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · 270d3eb4371b0d85 · report fixture: signature answer_dir:str/0/str · returned dict · level recorded 2026-09-24
load_questions in 17 papers mt_bench/common.py ran Apache-2.0 (permissive) · 88969cdc675c418e · report fixture: signature question_file:str/0/str, begin:int/0/int, end:int/0/int · returned list · level recorded 2026-09-24
revert mt_bench/compute_agreement.py ran fingerprinted Apache-2.0 (permissive) · 35801bf32affe123 · report fixture: signature vote:str/0/str · returned str · level recorded 2026-09-25
stabilized_log1pexp trainers.py ran fingerprinted Apache-2.0 (permissive) · 82acf8c8f8bc3e2b · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-24
load_model in 479 papers mt_bench/convert_dpo_trainer_file_to_huggingface.py unverified Apache-2.0 (permissive) · 3e9f2dd8699a6fff · report

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

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