Analyzing Toxicity in Deep Conversations: A Reddit Case Study