Abstract: Limiting undesired large language model (LLM) output, including: detecting that an output of a large language model (LLM) satisfies one or more conditions indicating that the output is undesirable; identifying a path in the large language model used to generate the output; and performing, based on the path, one or more remedial actions to modify how the path affects output by the LLM.

Abstract: Dynamic Combinatorial Test Design (CTD) modeling includes querying, by a processing device, a system under test (SUT) to be tested based on a CTD model. The processing device receives, based on the querying, system configuration information including one or more system value sets each corresponding to an attribute of the CTD model. The processing device modifies the CTD model based on the received configuration information.

Abstract: Dynamic CTD model tracking and optimization, including: identifying a difference in coverage between a CTD model and a dynamically generated CTD model generated by applying one or more test environment limitations to the CTD model; and maintaining a listing of uncovered value combinations.

Abstract: Real-time validation of microcode updates may include: executing a workload on at least one first processor having a first version of a microcode instruction set; executing the workload on at least one second processor having a second version of a microcode instruction set, wherein the second version of the microcode instruction set comprises an update of the first version of the microcode instruction set; and validating the second version of the microcode instruction set based on a comparison of output from executing the workload using the at least one first processor and output from executing the workload using the at least one second processor.

Abstract: A computer system, computer readable storage medium, and computer-implemented method for ranking and ordering software interrelationships. The method includes generating a ranking of software dependency combinations. The ranking is at least partially subject to one or more criteria. The method also includes ordering the ranked software dependency combinations.

Abstract: Tagging a corpus of tests based upon failure analysis. Access tags associated with test cases known to execute without error against an original corpus of source code. Access a test set associated with the accessed tags, the test set containing test cases. Access a new corpus of source code to be tested. Execute the new corpus of source code using test cases from the accessed test set. Access errors generated by execution of the new corpus of source code using the test cases. Determine locations in the new corpus of source code associated with the accessed errors. Execute natural language processing software to generate new tags associated with the locations in the new corpus of source code associated with the accessed errors.

Abstract: Method and apparatus for data backup. A first backup of a computing system is generated at a first time. A first confidence of compromise level of the computing system for the first time is generated. The first backup is stored along with metadata, where the metadata comprises the first confidence of compromise level of the computing system at the first time. In response to evaluating the first confidence of compromise level based on one or more backup criteria, a backup policy of the computing system is modified.

Abstract: A method of analyzing code is provided. The method includes generating an abstract representation of the code, identifying conditional statements in the abstract representation, populating a truth table for each conditional statement that has been identified with all possible outcomes of the conditional statement and assessing the truth table for each conditional statement to identify issues.

Abstract: Techniques for vector-based identification of software dependency relationships are described herein. An aspect includes determining a first dependency relationship value between a first code segment and a second code segment. Another aspect includes calculating a magnitude vector based on the first dependency relationship value and a second dependency relationship value corresponding to the first code segment and the second code segment. Another aspect includes determining a relationship score for the first code segment and the second code segment based on the magnitude vector and the first dependency relationship value.

Abstract: Techniques for vector-based identification of software dependency relationships are described herein. An aspect includes determining a first dependency relationship value between a first code segment and a second code segment. Another aspect includes calculating a magnitude vector based on the first dependency relationship value and a second dependency relationship value corresponding to the first code segment and the second code segment.

Abstract: Aspects of the invention include mutating each neural network of a portion of a first array of neural networks, wherein each neural network of the first array of neural networks is configured to select a respective sequence of test cases for testing a computing infrastructure. Causing each neural network of a second array of neural networks to select a respective sequence of test cases for testing the computing infrastructure. Generating a child neural network by performing a crossover operation between a mutated neural network of the portion of the first array and a neural network of the second array of neural networks, the child neural network generating a new sequence of test cases for testing the computing infrastructure.

Abstract: Techniques for subject matter expert identification for computer software are described herein. An aspect includes receiving a search string. Another aspect includes searching a plurality of library usage data files to determine a selected library corresponding to the search string. Another aspect includes determining a location of a library call corresponding to the selected library in source code. Another aspect includes determining a computer programmer associated with the determined location based on version control metadata corresponding to the source code. Another aspect includes identifying the determined computer programmer as a subject matter expert for the selected library.

Abstract: Techniques for subject matter expert identification for computer software are described herein. An aspect includes receiving a search string. Another aspect includes searching a plurality of library usage data files to determine a selected library corresponding to the search string. Another aspect includes determining a location of a library call corresponding to the selected library in source code. Another aspect includes determining a computer programmer associated with the determined location based on version control metadata corresponding to the source code. Another aspect includes identifying the determined computer programmer as a subject matter expert for the selected library.

Abstract: Techniques for an expertise score vector for software component management are described herein. An aspect includes determining a size and an amount of time corresponding to committed code contributed by a first developer to a first software component. Another aspect includes determining a time per unit of contribution based on the size and amount of time. Another aspect includes updating component mastery metrics corresponding to the first software component in an expertise score vector corresponding to the first developer based on the time per unit of contribution. Another aspect includes assigning the first developer to a developer tier based on the component mastery metrics. Another aspect includes assigning a work item corresponding to the first software component to the first developer based on the developer tier.

Abstract: Aspects of the invention include receiving a notification that a first logical grouping of code segments of a software application is being changed. Comparing the first logical grouping to a second logical grouping of code segments from a change history of the software application, where the first logical grouping and the second logical grouping both relate to resolving a common software application issue, and where the first logical grouping and the second logical grouping include at least one common code segment. Identifying a first code segment of the first logical grouping that is different from a second code segment of the second logical grouping. Determining that given the at least one code segment is changed a probability that the second code segment has been changed is greater than an threshold value. Sending a notification based on the difference in the probabilities being greater than a threshold value.

Abstract: A method of analyzing code is provided. The method includes generating an abstract representation of the code, identifying conditional statements in the abstract representation, populating a truth table for each conditional statement that has been identified with all possible outcomes of the conditional statement and assessing the truth table for each conditional statement to identify issues.

Abstract: A method of analyzing code is provided. The method includes generating an abstract representation of the code, identifying conditional statements in the abstract representation, populating a truth table for each conditional statement that has been identified with all possible outcomes of the conditional statement and assessing the truth table for each conditional statement to identify issues.

Abstract: A method for detecting and localizing a fault in a system under test (SUT) includes generating an initial set of test vectors that provides complete n-wise coverage of the reduced test space. The method further includes generating and executing an initial set of test cases to obtain a first set of execution results. The method further includes determining, based at least in part on the first set of execution results, that one or more test cases failed execution. The method further includes generating a set of new test cases from a selected failing test case. The method further includes executing the set of new test cases to obtain a second set of execution results. The method further includes detecting and localizing the fault based at least in part on the second set of execution results.

Abstract: Aspects of the invention include receiving, by a processor, a request to copy a code from a source file and receiving, by the processor, a request to paste the code into a destination file. Aspects also include creating, by the processor based at least in part on the request to paste the code, an entry in a database, the entry having an identification of the source file, an identification of the destination file, a location of the code in the source file, and a location of the code in the destination file.

Abstract: A method of analyzing code is provided. The method includes generating an abstract representation of the code, identifying conditional statements in the abstract representation, populating a truth table for each conditional statement that has been identified with all possible outcomes of the conditional statement and assessing the truth table for each conditional statement to identify issues.