---
license: cc-by-nc-4.0
language:
  - en
pretty_name: EchoSim Synthetic Compatibility Dialogues (Sample)
tags:
  - synthetic
  - dialogue
  - dating
  - conversational
  - personas
task_categories:
  - text-generation
size_categories:
  - n<1K
configs:
  - config_name: default
    data_files: synthetic_dating_dataset_sample.jsonl
---

# EchoSim Synthetic Compatibility Dialogues (Sample)

A **fully synthetic** corpus of AI-simulated first-contact dialogues between two dating
personas. Each record pairs two personas (MBTI, attachment style, interests, age range)
with a short conversation. Current release: **200 sessions**.

This is a public **research / schema sample** released by [EchoSim.AI](https://echosim.ai).
It is meant to illustrate the *shape* of the data used to study conversational compatibility —
not to expose any production system.

**Links:** [Official site](https://echosim.ai) · [GitHub specs & whitepaper](https://github.com/echo-sim/Echo-Sim-Official-Specs) · [Hugging Face org](https://huggingface.co/echo-sim) · [Demo Space](https://huggingface.co/spaces/echo-sim/echosim-first-contact-demo)

> ⚠️ **Synthetic data.** Every record is generated by LLM simulation from sampled persona
> attributes. There are **no real users, no real conversations, and no PII**. No persona
> corresponds to a real person.

## Why synthetic?

EchoSim simulates AI "Echoes" dating on a user's behalf. To share research artifacts without
touching real user data, we publish synthetic samples generated by the same *kind* of pipeline.
What is intentionally **not** included: the production matching algorithm, scoring logic, the
internal signal taxonomy, and any user-derived data.

## Dataset structure

One JSON object per line (`synthetic_dating_dataset_sample.jsonl`).

| Field | Type | Description |
|---|---|---|
| `session_id` | string | Stable id for the simulated session |
| `personas` | object | `user_A` / `user_B`, each with `mbti`, `attachment_style`, `interests[]`, `age_range` |
| `dialogue` | array | Ordered turns of `{role, message}` |

### Example

```json
{
  "session_id": "sim_001",
  "personas": {
    "user_A": {"mbti": "INTJ", "attachment_style": "Dismissive-Avoidant", "interests": ["Tech", "Reading", "Chess"], "age_range": "28-32"},
    "user_B": {"mbti": "ENFP", "attachment_style": "Secure", "interests": ["Travel", "Photography", "Live Music"], "age_range": "25-29"}
  },
  "dialogue": [{"role": "user_B", "message": "..."}]
}
```

## How it was generated

1. **Persona sampling** — combinatorial sampling across MBTI type, adult attachment style, an
   interest set, and a coarse age range.
2. **Dialogue generation** — an LLM role-plays both personas through a short first-contact
   exchange.

## Usage

```python
from datasets import load_dataset

ds = load_dataset("json", data_files="synthetic_dating_dataset_sample.jsonl", split="train")
print(ds[0])
```

## Limitations & intended use

- **Small corpus** (200 records). Intended for schema illustration, prototyping, and smoke tests rather than large-scale training.
- Content is English-only and may carry the biases of the generating LLM.
- Persona frameworks (MBTI, attachment style) are used as descriptive tags, not scientific claims.
- Do **not** use for real-world relationship, psychological, or hiring decisions.

## License

Released under **CC-BY-NC-4.0**. Attribution: EchoSim.AI (https://echosim.ai).