Skip to content

A LLM trained only on data from certain time periods to reduce modern bias

Python

Enhance README with language options and project info

Hayk Grigorian6036a4f0ec2e
diff --git a/README.md b/README.md
index 88fd6ce..60826be 100644
--- a/README.md
+++ b/README.md
@@ -116,9 +116,9 @@ Trained to 182K steps
### **v2**
-- 90GB of 1800-1875 London texts
+- 90GB(raw) of 1800-1875 London texts
- 136,344 documents
-- The full 90GB is not avalaible yet as it hasn't been tokenized but you can find a 15GB sample here: https://huggingface.co/datasets/haykgrigorian/TimeCapsuleLLM-London-1800-1875-v2-15GB
+- The full tokenized dataset is now avalaible here: https://huggingface.co/datasets/postgrammar/london-llm-1800
### Bias Stats
![Pronoun bias](https://github.com/haykgrigo3/TimeCapsuleLLM/blob/main/london_1800_1875_v2mini_eval1/pronoun_bias.png)
@@ -222,5 +222,6 @@ GPU: A100 SXM rented
+