Projects.
4 PROJECTS · 3 OPEN SOURCE · 6 MERGED UPSTREAM PRS
Research software I lead or co-author, and contributions to tools others depend on.
benchgap
Fills the empty cells of LLM leaderboards. Calibrates benchmarks against each other on the models measured on both, predicts each missing score from that model's own results, and attaches a cross-validated error and confidence level to every estimate. Data refreshed daily from BenchLM, with a public OpenAPI 3.1 JSON API. Built with Roland Hollós.
- Python
- Curve fitting
- LLM evaluation
- PHP
Distribird
Multi-agent LangGraph pipeline that turns scientific literature into informative Bayesian priors. Searches Semantic Scholar and OpenAlex in parallel, extracts reported numerical values, and fits a probability distribution via AIC model selection. Built with HUN-REN AI1Science.
- Python
- LangGraph
- LLM agents
- Bayesian inference
Syxplain
Neurosymbolic web platform that turns a table of numbers into a readable law. An LLM agent steers PySR symbolic regression on a Slurm HPC cluster, interprets the whole Pareto front, and grounds every explanation in real literature through deep research. Built with HUN-REN AI1Science.
- Python
- FastAPI
- React
- PySR
- Slurm
SpotBugs
6 merged pull requests to the widely-used Java static-analysis tool, including two new CERT Java detectors with full test cases: SER11-J (prevent overwriting of externalizable objects) and VNA04-J (ensure calls to chained methods are atomic). The atomicity detector connects to the published Acta Cybernetica paper on multithreading atomicity.
- Java
- Static analysis
- CERT Java
- AST visitors
- #3037Increase Version class log levelRestored the output of spotbugs -version, which an earlier switch to info-level logging had silenced.merged2024-07-11+9 / −2
- #2986Add 'VNA04-J. Ensure that calls to chained methods are atomic' detectorNew CERT detector for chained method calls on shared objects that are not atomic as a whole, with test cases.merged2024-10-09+830 / −27
- #2919Add 'VNA03-J. Sequence of calls on a synchronized abstraction may not be atomic' detectorCompanion detector for sequences of individually synchronized calls; under review.open2024-04-03+1897 / −47
- #2823Remove AppleExtension libraryDropped the obsolete macOS adapter after verifying the GUI menus still work on current macOS.merged2024-01-20+1 / −216
- #2762Fix false positive in SE_PREVENT_EXT_OBJ_OVERWRITEFixed a false positive reported against the SER11-J detector (#2750).merged2023-12-12+315 / −13
- #2611Update Eclipse plugin release number to 4.29.0Release housekeeping for the Eclipse plugin (#2280).merged2023-10-07+1 / −1
- #2426SER11-J. Prevent overwriting of externalizable objectsNew CERT detector for readExternal implementations that can overwrite an already initialized object.merged2023-06-09+335 / −5