2 Gram model in C#

NGram.cs

using System;
using System.IO;
using System.Collections.Generic;
using System.Linq;

namespace generating_text
{
    class Model 
    {
        // A map of Word => {Words}.
        // Note that words are duplicated according to the frequency they occur
        private Dictionary<string, List<string>> _languageModel;                
        private Model(Dictionary<string,List<string>> languageModel)
        {
            _languageModel = languageModel;
        }

        public static Model Build(IEnumerable<string> words) 
        {
            var m = new Dictionary<string, List<string>>();          
            foreach (var (fst,snd) in words.Zip(words.Skip(1)))
            {
                m[fst] = m.GetValueOrDefault(fst, new List<string>());
                m[fst].Add(snd);
            }

            return new Model(m);
         }

        private Random random = new Random();

        private string Next(string seed)
        {
            if (!_languageModel.TryGetValue(seed, out var nextWords))
                return null;

            return nextWords[random.Next(nextWords.Count())];
        }

        public IEnumerable<string> Generate(string word)
        {
            do {
                yield return word;
            } while (null != (word = Next(word)));
        }
    }

    class Program
    {
        static void Main(string[] args)
        {
            // Read from standard in
            var input = args.Length >= 1 ? File.ReadAllText(args[0]).ToLower() : Console.In.ReadToEnd().ToLower();

            // Split by whitespace and trim out blank words
            var words = input.Split(null).Where(x => x.Trim().Length != 0);

            // Create a language model
            var model = Model.Build(words);

            Console.Out.WriteLine(string.Join(" ", model.Generate("the").Take(100).ToArray()));
        }
    }
}