Verification Witnesses from Verification Tools (SV-COMP 2022)
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
SV-COMP 2022 Verification Witnesses This file describes the contents of an archive of the 11th Competition on Software Verification (SV-COMP 2022).<br> https://sv-comp.sosy-lab.org/2022/ The competition was run by Dirk Beyer, LMU Munich, Germany.<br> More information is available in the following article:<br> Dirk Beyer. <em>Progress on Software Verification: SV-COMP 2022.</em> In Proceedings of the 28th International Conference on Tools and Algorithms for the Construction and Analysis of Systems (TACAS 2022, Munich, April 2 - 7), 2022. Springer. Copyright (C) Dirk Beyer<br> https://www.sosy-lab.org/people/beyer/ SPDX-License-Identifier: CC-BY-4.0<br> https://spdx.org/licenses/CC-BY-4.0.html Contents <code>LICENSE.txt</code>: specifies the license <code>README.txt</code>: this file <code>witnessFileByHash/</code>: This directory contains verification witnesses. Each verification witness in this directory is stored in a file whose name is the SHA2 256-bit hash of its contents followed by the filename extension .graphml. The format of each verification witness is described on the format web page: https://github.com/sosy-lab/sv-witnesses/ A verification witness contains also metadata in order to relate it to the verification task for which it was produced. <code>witnessInfoByHash/</code>: This directory contains for each verification witness in directory witnessFileByHash/ a record in JSON format (also using the SHA2 256-bit hash of the witness as filename, with .json as filename extension) that contains the meta data. <code>witnessListByProgramHashJSON/</code>: For convenient access to all verification witnesses for a certain program, this directory represents a function that maps each program (via its SHA2256-bit hash) to a set of verification witnesses (JSON records for verification witnesses as described above) that the verification tools have produced for that program. For each program for which verification witnesses exist, the directory contains a JSON file (using the SHA2 256-bit hash of the program as filename, with .json as filename extension) that contains all JSON records for verification witnesses for that program. The data structure is described in the following article:<br> Dirk Beyer. <em>A Data Set of Program Invariants and Error Paths.</em> In Proceedings of the 2019 IEEE/ACM 16th International Conference on Mining Software Repositories (MSR 2019, Montreal, Canada, May 26-27), pages 111-115, 2019. IEEE.<br> https://doi.org/10.1109/MSR.2019.00026 Other Archives Overview over archives from SV-COMP 2022 that are available at Zenodo: https://doi.org/10.5281/zenodo.5838498 Verification Witnesses from SV-COMP 2022 Verification Tools. Witness store (containing the generated verification witnesses) https://doi.org/10.5281/zenodo.5831008 Results of the 11th Intl. Competition on Software Verification (SV-COMP 2022). Results (XML result files, log files, file mappings, HTML tables) https://doi.org/10.5281/zenodo.5831003 SV-Benchmarks: Benchmark Set of SV-COMP 2022 and Test-Comp 2022. Verification tasks, version svcomp22 https://doi.org/10.5281/zenodo.5720267 BenchExec, version 3.10. Benchmarking framework All benchmarks were executed for SV-COMP 2022 https://sv-comp.sosy-lab.org/2022/<br> by Dirk Beyer, LMU Munich, based on the following components: https://gitlab.com/sosy-lab/sv-comp/archives-2022 svcomp22 a6b18082 https://gitlab.com/sosy-lab/benchmarking/sv-benchmarks svcomp22 ad265d07 https://gitlab.com/sosy-lab/sv-comp/bench-defs svcomp22 0332884a https://gitlab.com/sosy-lab/software/benchexec 3.10 4e8716bd https://gitlab.com/sosy-lab/benchmarking/competition-scripts svcomp22 3c959671 https://github.com/sosy-lab/sv-witnesses svcomp22 e4695d2b Contact Feel free to contact me in case of questions: https://www.sosy-lab.org/people/beyer/
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.004 | 0.000 |
| Scholarly communication | 0.003 | 0.001 |
| Open science | 0.007 | 0.005 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.091 | 0.006 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it