Papers › How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources

How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources

7 Jun 2023NeurIPS 2023 11arXiv:2306.04751archive 2025-07-28

Yizhong Wang, Hamish Ivison, Pradeep Dasigi, Jack Hessel, Tushar Khot, Khyathi Raghavi Chandu, David Wadden, Kelsey MacMillan, Noah A. Smith, Iz Beltagy, Hannaneh Hajishirzi

In this work we explore recent advances in instruction-tuning language models on a range of open instruction-following datasets. Despite recent claims that open models can be on par with state-of-the-art proprietary models, these claims are often accompanied by limited evaluation, making it difficult to compare models across the board and determine the utility of various resources. We provide a large set of instruction-tuned models from 6.7B to 65B parameters in size, trained on 12 instruction datasets ranging from manually curated (e.g., OpenAssistant) to synthetic and distilled (e.g., Alpaca) and systematically evaluate them on their factual knowledge, reasoning, multilinguality, coding, and open-ended instruction following abilities through a collection of automatic, model-based, and human-based metrics. We further introduce Tülu, our best performing instruction-tuned model suite finetuned on a combination of high-quality open resources. Our experiments show that different instruction-tuning datasets can uncover or enhance specific skills, while no single dataset (or combination) provides the best performance across all evaluations. Interestingly, we find that model and human preference-based evaluations fail to reflect differences in model capabilities exposed by benchmark-based evaluations, suggesting the need for the type of systemic evaluation performed in this work. Our evaluations show that the best model in any given evaluation reaches on average 87% of ChatGPT performance, and 73% of GPT-4 performance, suggesting that further investment in building better base models and instruction-tuning data is required to close the gap. We release our instruction-tuned models, including a fully finetuned 65B Tülu, along with our code, data, and evaluation framework at https://github.com/allenai/open-instruct to facilitate future research.

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 12 code samples harvested from 4 repositories linked to this paper: 6 ran (of which 1 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 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: official repository: 8 samples from 1 repository, 5 ran (2 with no instrument failure; 3 where Syntology's instrument failed); community (archive-listed): 4 samples from 3 repositories, 1 ran (0 with no instrument failure; 1 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 2026-09-22 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 1 of the 6 samples that ran has no recorded date.

allenai/open-instruct officialmentioned in papermentioned on GitHubpytorchSyntology: 8 samples, 5 ran (3 instrument-failed) · 3 unverified report
Forence1999/PRoLoRA mentioned on GitHubpytorchSyntology: 1 sample, 0 ran · 1 unverified report
john-hewitt/implicit-ins mentioned on GitHubpytorchSyntology: 1 sample, 0 ran · 1 unverified report
schwartz-lab-NLP/label-bias mentioned on GitHubpytorchSyntology: 2 samples, 1 ran (1 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

12 samples harvested: 6 ran (of which 1 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 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 has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · honoured contract
2ran · our draft was wrong
2ran; the contract check could not run
1ran
6unverified

Licence: 2 of the 12 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 4 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.

allenai/open-instruct official repository 8 samples, 5 ran (3 instrument-failed) · 3 unverified
entropy_from_logits in 13 papers open_instruct/model_utils.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · 3ca03473e153b197 · report fixture: signature logits:Tensor/2/float · returned Tensor · level recorded, date not recorded
AssistantSpanDerivationError open_instruct/dataset_transformation.py ran Apache-2.0 (permissive) · 7c7fbc84c32888c3 · report fixture: returned AssistantSpanDerivationError · level recorded 2026-09-22
_assistant_token_spans_from_prefix_lengths open_instruct/dataset_transformation.py ran; the contract check could not run (RAISES) Apache-2.0 (permissive) · d38e226c6b168c4a · report fixture: signature messages:list/1/dict, tokenizer:MockTokenizer/0/?, tools:NoneType/0/?, max_seq_length:int/0/int, trainable_indices:list/1/int · returned list · level recorded 2026-09-22
_trainable_assistant_indices open_instruct/dataset_transformation.py ran · honoured contract Apache-2.0 (permissive) · 4baec48bd245f848 · report fixture: signature messages:list/1/dict, last_turn_only:bool/0/bool · returned list · level recorded 2026-09-22
_was_truncated open_instruct/dataset_transformation.py ran; the contract check could not run (RAISES) Apache-2.0 (permissive) · 29d4865962de784a · report fixture: signature offsets:list/1/list, rendered:str/0/str, n_tokens:int/0/int, max_seq_length:int/0/int · returned bool · level recorded 2026-09-22
_tokenize_tulu_sft_with_assistant_labels open_instruct/dataset_transformation.py unverified Apache-2.0 (permissive) · b027fd7ccc6d4be9 · report
_verify_assistant_spans_cover_content open_instruct/dataset_transformation.py unverified Apache-2.0 (permissive) · cfe06104ea9d81d9 · report
take_and_check_valid_input scripts/add_metadata.py unverified Apache-2.0 (permissive) · b252d3f8f3f983ba · report
Forence1999/PRoLoRA community (archive-listed) 1 sample, 0 ran · 1 unverified
LoraLayer_update_layer modules/chunkwise_sharing_LoRA.py unverified Apache-2.0 (permissive) · 370f948ce836365d · report
john-hewitt/implicit-ins community (archive-listed) 1 sample, 0 ran · 1 unverified
CombinedCausalLM open_instruct/combined_model.py unverified Apache-2.0 (permissive) · a7124b520aa932e3 · report
schwartz-lab-NLP/label-bias community (archive-listed) 2 samples, 1 ran (1 instrument-failed) · 1 unverified
find_all_linear_names in 50 papers src/superni/utils/lora_utils.py ran · our draft was wrong no licence file found · pointer only · bea710de6ec7c18d · report fixture: signature model:callable/0/? · returned set · level recorded 2026-09-22
prepare_peft_model src/superni/utils/lora_utils.py unverified no licence file found · pointer only · d5011eb064a2523d · report

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.

4 assemblies on this paper: 4 harvested, 0 generated, listed in 1 groups (siblings lifted in one batch form one group).

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

harvested Assembly for TÜLU (graph method: TÜLU). 2 bricks; 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) · 1 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. prepare_peft_model entry · level 0 · unverified · d5011eb064a2523d
  2. find_all_linear_names helper · level 1 · ran · our draft was wrong · bea710de6ec7c18d

harvested Assembly for TÜLU (graph method: TÜLU). 1 brick; individually on a fixture: 0 ran · 1 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. LoraLayer_update_layer entry · level 0 · unverified · 370f948ce836365d

harvested Assembly for TÜLU (graph method: TÜLU). 6 bricks; individually on a fixture: 4 ran (of which 1 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. _tokenize_tulu_sft_with_assistant_labels entry · level 0 · unverified · b027fd7ccc6d4be9
  2. AssistantSpanDerivationError helper · level 1 · ran · 7c7fbc84c32888c3
  3. _assistant_token_spans_from_prefix_lengths helper · level 1 · ran; the contract check could not run (RAISES) · d38e226c6b168c4a
  4. _trainable_assistant_indices helper · level 1 · ran · honoured contract · 4baec48bd245f848
  5. _verify_assistant_spans_cover_content helper · level 0 · unverified · cfe06104ea9d81d9
  6. _was_truncated helper · level 1 · ran; the contract check could not run (RAISES) · 29d4865962de784a

harvested Assembly for TÜLU (graph method: TÜLU). 1 brick; individually on a fixture: 0 ran · 1 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. CombinedCausalLM entry · level 0 · unverified · a7124b520aa932e3

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="2306.04751") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Instruction Following

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Absolute Position EncodingsAdamAttentionBASEBPEDense ConnectionsDropoutGPT-4Label SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxTransformer

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