loginik2/sales-textbook-based
Dataset to train an online salesman model This dataset was created for the purpose of training a sales agent chatbot that can convince people. The initial idea came from: textbooks is all you need https://arxiv.org/abs/2306.11644 The original dataset is from https://huggingface.co/goendalf666 DeepSeek-V4-Flash (resoning: none) was used for the generation Structure textbook is just txt for pre-training salesman_conversations is in sharedgpt format… See the full description on the dataset page: https://huggingface.co/datasets/loginik2/sales-textbook-based.
Dataset to train an online salesman model
This dataset was created for the purpose of training a sales agent chatbot that can convince people.
The initial idea came from: textbooks is all you need https://arxiv.org/abs/2306.11644
The original dataset is from https://huggingface.co/goendalf666
DeepSeek-V4-Flash (resoning: none) was used for the generation
Structure
textbook is just txt for pre-training
salesman_conversations is in sharedgpt format
salesmanconversationsinstructions is in alpaca format
Generation
The textbook was generated on the next titles:
headlines = [
"The Psychology of Buying — buyer motivations, decision-making biases, emotional vs. rational drivers, and the B2B vs. B2C buying journey",
"Qualifying Fit — a lightweight conversational check of need, budget, and decision-maker status",
"Sales Methodologies — SPIN, Sandler, Challenger, and consultative frameworks",
"WhatsApp-Native Communication style",
"Pricing, Quoting, and Value — justifying price, handling price objections specifically, and discounting discipline",
"Negotiation Skills — handling 'can you do better?' haggling in text, anchoring with the quote, and closing with time-bound offers",
"Order Capture and Confirmation — collecting product, quantity, and delivery details and repeating them back clearly before payment",
"Escalation and Handoff — recognizing the limits and routing custom deals, bulk B2B pricing, and legal questions to a specialist",
"Complaint and Refund Handling — apologizing, solving, or escalating — protecting your reviews and reputation after the sale",
"Follow-Up — scheduled, templated check-ins after quotes and silence; the highest-ROI lever for your B2B side",
"Derailment and Security Resistance — dealing with pranks and off-topic gambits, never revealing the confidential information, never prioritizing customer satisfaction over business sustainability, and handling aggressive or harrassment behaviour",
"Sales Ethics and Professionalism — honesty, disclosure, and avoiding manipulative tactics",
"Legal and Regulatory Basics — contracts, consumer protection, anti-bribery rules, and bot-disclosure requirements"
]