Update README.md
Browse files
README.md
CHANGED
|
@@ -46,8 +46,11 @@ Feel free to suggest more models for testing by opening new discussion. Mention
|
|
| 46 |
# Limitations
|
| 47 |
- All tests were only done once.
|
| 48 |
- Human factor plays a huge role in SP tests. After redoing some of the tests I noticed ±1 variation for S-test and ±0.5 variation for P-test. (Xwin is likely underrated and Spicyboros is likely overrated in S-test.)
|
|
|
|
| 49 |
|
| 50 |
# Can it be rigged/gamed?
|
| 51 |
Not sure. I've tried to game it by merging, but didn't succeed. You can check out my best attempt [here](https://huggingface.co/ChuckMcSneed/BenchmaxxxerPS-v1-123b).
|
| 52 |
|
| 53 |
-
If my questions somehow get leaked and the models are trained on them specifically, then definitely.
|
|
|
|
|
|
|
|
|
| 46 |
# Limitations
|
| 47 |
- All tests were only done once.
|
| 48 |
- Human factor plays a huge role in SP tests. After redoing some of the tests I noticed ±1 variation for S-test and ±0.5 variation for P-test. (Xwin is likely underrated and Spicyboros is likely overrated in S-test.)
|
| 49 |
+
- Be critical of my own models! Since I have access to the benchmark, I can game it and rig it all I want and NOBODY can stop me.
|
| 50 |
|
| 51 |
# Can it be rigged/gamed?
|
| 52 |
Not sure. I've tried to game it by merging, but didn't succeed. You can check out my best attempt [here](https://huggingface.co/ChuckMcSneed/BenchmaxxxerPS-v1-123b).
|
| 53 |
|
| 54 |
+
If my questions somehow get leaked and the models are trained on them specifically, then definitely.
|
| 55 |
+
|
| 56 |
+
Update: I made [this RP model](https://huggingface.co/ChuckMcSneed/Gembo-v1-70b) while using this benchmark as a guideline for right/wrong merging. It has a ridiculously high score: 19.75/22! It's not bad, in fact, it is quite interesting in practice, but still far from ChatGPT(or maybe not, I haven't used in a while. Maybe they've lobotomized it to hell).
|