Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -45,7 +45,6 @@ A simple context interference evaluation.
|
|
| 45 |
I welcome discussions and collaborations with research labs and model teams on **mechanistic interpretability—especially how LLMs represent, update, and lose information—rigorous model and agent evaluation, and why models fail, not only whether they fail**. If you are interested in PI-LLM, benchmark comparisons, or evaluating your models or agents, please reach out.
|
| 46 |
|
| 47 |
**Email:** [wangchupei@hotmail.com](mailto:wangchupei@hotmail.com)
|
| 48 |
-
**WeChat:** `wangchupei` — please mention **“PI-LLM”** in your friend request.
|
| 49 |
|
| 50 |
I usually reply within 36 hours.
|
| 51 |
|
|
|
|
| 45 |
I welcome discussions and collaborations with research labs and model teams on **mechanistic interpretability—especially how LLMs represent, update, and lose information—rigorous model and agent evaluation, and why models fail, not only whether they fail**. If you are interested in PI-LLM, benchmark comparisons, or evaluating your models or agents, please reach out.
|
| 46 |
|
| 47 |
**Email:** [wangchupei@hotmail.com](mailto:wangchupei@hotmail.com)
|
|
|
|
| 48 |
|
| 49 |
I usually reply within 36 hours.
|
| 50 |
|