Papers › Exploring the impact of low-rank adaptation on the performance, efficiency, and...

Exploring the impact of low-rank adaptation on the performance, efficiency, and regularization of RLHF

16 Sep 2023arXiv:2309.09055archive 2025-07-28

Simeng Sun, Dhawal Gupta, Mohit Iyyer

During the last stage of RLHF, a large language model is aligned to human intents via PPO training, a process that generally requires large-scale computational resources. In this technical report, we empirically investigate an efficient implementation of RLHF using low-rank adaptation (LoRA), which allows us to align the LLaMA 7B checkpoint on the Alpaca dataset using only two A100 GPUs instead of the eight required for full model fine-tuning. Despite tuning only 0.2% of LLaMA 7B's parameters, our implementation achieves better performance than the publicly-released AlpacaFarm checkpoint with full model fine-tuning. Next, we analyze several configurations of our LoRA-based PPO implementation, varying the form of the KL regularization term in the training objective. We find that (1) removing this penalty term does not harm performance on the AlpacaFarm evaluation set under our LoRA setup; (2) other regularizers, such as Jensen-Shannon divergence, lead to improved performance; and (3) while PPO training negatively impacts the factuality of model-generated responses, training with LoRA largely mitigates this effect. We release our code and pretrained checkpoints to facilitate future research on more efficient RLHF.

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 13 code samples harvested from 1 repository linked to this paper: 12 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 3 where Syntology's instrument failed) · 1 unverified. Where Syntology's instrument failed: 3 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 13 samples from 1 repository, 12 ran (9 with no instrument failure; 3 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 on 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

simengsun/alpaca_farm_lora officialmentioned in paperjaxApache-2.0Syntology: 13 samples, 12 ran (3 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

13 samples harvested: 12 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 3 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.

3ran · our draft was wrong
9ran
1unverified

Licence: 0 of the 13 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 simengsun/alpaca_farm_lora. “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.

simengsun/alpaca_farm_lora official repository 13 samples, 12 ran (3 instrument-failed) · 1 unverified
compute_reward_modeling_metrics in 5 papers alpaca_farm/src/alpaca_farm/reward_modeling_trainer.py ran · our draft was wrong Apache-2.0 (permissive) · cca45bb9a111ab77 · report fixture: signature eval_prediction:EvalPrediction/0/? · returned dict · level recorded 2026-09-24
create_scheduler in 17 papers alpaca_farm/src/alpaca_farm/trainer_utils.py ran Apache-2.0 (permissive) · 42596ac916e883a5 · report fixture: signature args:Args/0/?, optimizer:Adam/0/?, lr_scheduler:NoneType/0/?, num_training_steps:int/0/int · returned LambdaLR · level recorded 2026-09-24
flatten_dict in 72 papers alpaca_farm/src/alpaca_farm/common.py ran Apache-2.0 (permissive) · f5ff77db7757bc92 · report fixture: signature nested:dict/0/?, sep:str/0/str, postprocess_fn:callable/0/? · returned dict · level recorded 2026-09-24
format_output in 8 papers alpaca_farm/src/alpaca_farm/data_preprocessor.py ran · our draft was wrong Apache-2.0 (permissive) · 46af3eb303562401 · report fixture: signature example:dict/0/?, eos_token:str/0/str, output_key:str/0/str · returned str · level recorded 2026-09-24
format_prompt in 45 papers alpaca_farm/src/alpaca_farm/data_preprocessor.py ran Apache-2.0 (permissive) · 5ef0d2db8e674cd5 · report fixture: signature example:dict/0/?, prompt_dict:dict/0/? · returned str · level recorded 2026-09-24
format_prompt_with_data_frame alpaca_farm/src/alpaca_farm/data_preprocessor.py ran Apache-2.0 (permissive) · 27d54a5b1d488ca0 · report fixture: signature df:DataFrame/0/?, prompt_dict:dict/0/?, df_postprocessor:callable/0/?, return_dict:bool/0/bool · returned tuple · level recorded 2026-09-24
get_logger in 383 papers alpaca_farm/src/alpaca_farm/logging.py ran Apache-2.0 (permissive) · 043cf4a6fd60480b · report fixture: signature name:str/0/str, log_level:str/0/str · returned MultiProcessAdapter · level recorded 2026-09-24
let_model_save_mem_when_zero_grad alpaca_farm/src/alpaca_farm/common.py ran · our draft was wrong Apache-2.0 (permissive) · 9a1871dd5c967ac1 · report fixture: signature model:callable/0/? · returned SimpleModel · level recorded 2026-09-24
make_rl_data_module alpaca_farm/src/alpaca_farm/data_utils.py ran Apache-2.0 (permissive) · 0612fd8d6b151bf4 · report fixture: returned dict · level recorded 2026-09-24
requires_chatml alpaca_farm/src/alpaca_farm/openai_utils.py ran Apache-2.0 (permissive) · 9bd28dd37074fcb1 · report fixture: signature model:str/0/str · returned bool · level recorded 2026-09-24
setup in 24 papers alpaca_farm/src/alpaca_farm/distributed_utils.py ran Apache-2.0 (permissive) · 80675b2c65496f66 · report fixture: signature rank:NoneType/0/?, world_size:NoneType/0/? · returned tuple · level recorded 2026-09-24
string_to_dict alpaca_farm/src/alpaca_farm/openai_utils.py ran fingerprinted Apache-2.0 (permissive) · 58d5d20497232727 · report fixture: signature to_convert:str/0/str · returned dict · level recorded 2026-09-24
make_binary_reward_modeling_data_module alpaca_farm/src/alpaca_farm/data_utils.py unverified Apache-2.0 (permissive) · 5451705c3c9a6d15 · report

Tasks

Language ModellingLarge Language Model

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

ALIGNEntropy RegularizationPPO

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