Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- text-generation
|
| 4 |
+
- instruction-following
|
| 5 |
+
- chat
|
| 6 |
+
- timeslast-ai
|
| 7 |
+
license: apache-2.0
|
| 8 |
+
language:
|
| 9 |
+
- en
|
| 10 |
+
base_model:
|
| 11 |
+
- TimesLast/TimesLastAI-4B
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
# TimesLast AI - 4B (Beta)
|
| 15 |
+
|
| 16 |
+
**TimesLast AI** is a 4-billion parameter, general-purpose conversational model developed by TimesLast.
|
| 17 |
+
|
| 18 |
+
This model was created to serve as a more direct, engaging, and less sycophantic alternative to contemporary AI assistants. Its core personality is witty, sarcastic and helpful, allowing for more natural and candid interactions.
|
| 19 |
+
|
| 20 |
+
## Model Description
|
| 21 |
+
|
| 22 |
+
- **Developed by:** TimesLast
|
| 23 |
+
- **Model Type:** Generative Language Model
|
| 24 |
+
- **License:** Apache 2.0
|
| 25 |
+
- **Primary Use:** General-purpose chat, complex instruction following, multi-language tasks, and coding assistance.
|
| 26 |
+
|
| 27 |
+
## Key Features
|
| 28 |
+
|
| 29 |
+
* **Distinct Personality:** Designed to be witty, sarcastic, and direct, avoiding the overly sanitized and obsequious language common in other models.
|
| 30 |
+
* **General Purpose:** While excelling at conversational banter, the model is highly capable in a wide range of domains including complex coding, technical explanations, creative writing, and historical analysis.
|
| 31 |
+
* **Multilingual Capabilities:** Includes training data in English and other languages making it effective for a variety of regional tasks.
|
| 32 |
+
* **Efficiency:** Delivers powerful results in a compact, 4-billion parameter package.
|
| 33 |
+
|
| 34 |
+
### Developer's Note
|
| 35 |
+
|
| 36 |
+
Due to the unique distillation process used in its creation, the model may occasionally state that it has 1 trillion parameters. This is a known quirk of its training and a testament to its ambitious design, not a reflection of its actual 4B architecture.
|