diff --git a/fault-readme.md b/fault-readme.md
index 8733b35..8d4d801 100644
--- a/fault-readme.md
+++ b/fault-readme.md
@@ -162,7 +162,7 @@ Currently, four fault models are implemented: "set1", "set0", "toggle", and "ove
 "set1" replaces all bits that are defined by fault_mask with 1 (OR), "set0" with 0 (NAND) and "toggle" toggles these bits (XOR).
 
 Overwrite allows to set a location ( instruction, memory or register) to a specific value. With it an instruction could be rewritten to a nop instruction on the fly. This is done with the overwrite example in fault.json file. The value to be written is defined by "fault_mask" parameter. It is written to "fault_address", beginning with the lowest byte and incrementing. It therefore is little-endian.
-num_bytes is used to determen how many bytes should be overwritten. Up to a maximum of 16 bytes can currently be overwritten with the help of this model.
+num_bytes is used to determine how many bytes should be overwritten. Up to a maximum of 16 bytes can currently be overwritten with the help of this model.
 
 ##### fault_lifespan
 Defines the number of instructions a fault is valid. If fault_lifespan is [0] it means the fault is permanent. If it is a positive number, fault_lifespan defines how many instructions the fault remains in the system. After this number of instructions, the fault is reverted. This has the potential to behave in a strange way, if the system changed the content at the respective address while the temporary fault was still active. It will only revert flipped bits back to the original state. Temporary faults are, thus, only recommended for flash memory, since usually in case of flash, less write accesses are occurring.