Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Set-LLM: A Permutation-Invariant LLM
Authors: Beni Egressy, Jan Stühmer
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We provide a theoretical proof of invariance and demonstrate through experiments that Set-LLM can be trained effectively, achieving comparable or improved performance and maintaining the runtime of the original model, while altogether eliminating order sensitivity. |
| Researcher Affiliation | Academia | 1Heidelberg Institute for Theoretical Studies 2IAR, Karlsruhe Institute of Technology |
| Pseudocode | Yes | An example is provided in Figure 3, along with pseudocode in Section B. Algorithm 1 Python-like pseudocode for calculating Set Position Encoding (Set PE) positions. |
| Open Source Code | Yes | All code is available under open licenses at https://github.com/hits-mli/set-llm. |
| Open Datasets | Yes | We get all the datasets from Hugging Face Datasets [24]. Table 5 provides metadata and Table 6 provides licensing details for each dataset. |
| Dataset Splits | Yes | We use the original train-validation-test split from [22]: 60% training, 10% validation, and 30% test data. |
| Hardware Specification | Yes | All models were trained and evaluated on Nvidia H200 GPUs on an internal cluster. |
| Software Dependencies | No | The paper mentions the Hugging Face PEFT library [30] and refers to bfloat16/32-bit floating point precision, implying the use of deep learning frameworks, but does not provide specific version numbers for any software components or libraries. |
| Experiment Setup | Yes | Details about the hyperparameter settings can be found in Section D.4. We include hyperparameter values in Section D.4 of the supplementary material and will also include these along with scripts to run specific experiments in the code release. |