Since greybox fuzzers (e.g., AFL++, libfuzzer) generate test inputs by random mutations guided toward high code coverage, it is challenging for human developers to analyze fuzzed crashing inputs to understand failure-inducing condition and locate buggy code elements for debugging. To address this challenge, we present DiffMin, a technique that refines a fuzzed crashing input to gain greater similarities to given passing inputs to help human developers better diagnose the fault. DiffMin iteratively applies edit actions to transform the crashing input while preserving the crash behavior. Our pilot study using four buggy programs from the Magma benchmark demonstrates that DiffMin effectively minimizes the differences between crashing and passing inputs while enhancing the accuracy of spectrum-based fault localization, highlighting its potential as a valuable pre-debugging step after greybox fuzzing.