Device-free human activity recognition has become a topic of much interest in recent years. While there is much effort into course-grained human activity recognition, the recognition of fine-grained human activities is still a research challenge. This paper presents a human activity recognition system to recognize fine-grained human activities in the dining context. We utilized an ESP32 microcontroller as a WiFi transceiver and gathered reflected Channel State Information (CSI) and Received Signal Strength Indicator (RSSI) from participants doing basic knife activities that are common in cooking. By a combination of CSI and RSSI at the feature level and using a Support Vector Machine (SVM) classifier, we achieved 74.2% accuracy in recognition of three fine-grained knife activities (chop, French cut, and slice).