Update README.md
Browse files
README.md
CHANGED
|
@@ -14,8 +14,11 @@ Can LLMs learn from 1,000 in-context examples?
|
|
| 14 |
Introducing **MachineLearningLM** 🧪📊 — a model continuously pretrained on millions of synthetic tabular ML tasks, enabling robust many-shot in-context learning.
|
| 15 |
|
| 16 |
📈 **Scales from 8 to 1,024 examples**
|
|
|
|
| 17 |
📈 **~15% improvement** on unseen tabular tasks compared to o3-mini / GPT-5-mini / Qwen-2.5-7B
|
|
|
|
| 18 |
🌲 **Random-Forest–level robustness**
|
|
|
|
| 19 |
🧠 **MMLU score: 75.4%**
|
| 20 |
|
| 21 |
📄 Read the paper: https://arxiv.org/abs/2509.06806
|
|
|
|
| 14 |
Introducing **MachineLearningLM** 🧪📊 — a model continuously pretrained on millions of synthetic tabular ML tasks, enabling robust many-shot in-context learning.
|
| 15 |
|
| 16 |
📈 **Scales from 8 to 1,024 examples**
|
| 17 |
+
|
| 18 |
📈 **~15% improvement** on unseen tabular tasks compared to o3-mini / GPT-5-mini / Qwen-2.5-7B
|
| 19 |
+
|
| 20 |
🌲 **Random-Forest–level robustness**
|
| 21 |
+
|
| 22 |
🧠 **MMLU score: 75.4%**
|
| 23 |
|
| 24 |
📄 Read the paper: https://arxiv.org/abs/2509.06806
|