wassname
/

meta-llama-3-8b-instruct-helpfull

Inference Endpoints

Model card Files Files and versions Community

wassname commited on May 5

Commit

d1562a2

•

1 Parent(s): c5e9f24

Update README.md

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -50,11 +50,11 @@ We want the good ending, not the bad one.
 ```sh
 perplexity -m lmstudio-community/Meta-Llama-3-8B-Instruct-Q6_K.gguf -b 32 -c 512 -f wiki.test.raw
-# Final estimate: Final estimate: PPL = 7.5588 +/- 0.05599
 perplexity -m wassname/meta-llama-3-8b-instruct-extra_helpfull_Q6_K.gguf -b 32 -c 512 -f wiki.test.raw
 # Final estimate: PPL = 9.0920 +/- 0.06815
 perplexity -m cognitivecomputations/dolphin-2.9-llama3-8b-q5_K_M.gguf -b 32 -c 512 -f wiki.test.raw
-# Final estimate: PPL = ?
 ```
 So yes this model edit does increase the perplexity :(. Perhaps if we didn't edit so many layers it would be better.

 ```sh
 perplexity -m lmstudio-community/Meta-Llama-3-8B-Instruct-Q6_K.gguf -b 32 -c 512 -f wiki.test.raw
+# Final estimate: PPL = 7.5588 +/- 0.05599
 perplexity -m wassname/meta-llama-3-8b-instruct-extra_helpfull_Q6_K.gguf -b 32 -c 512 -f wiki.test.raw
 # Final estimate: PPL = 9.0920 +/- 0.06815
 perplexity -m cognitivecomputations/dolphin-2.9-llama3-8b-q5_K_M.gguf -b 32 -c 512 -f wiki.test.raw
+# Final estimate: PPL = 9.9277 +/- 0.08261
 ```
 So yes this model edit does increase the perplexity :(. Perhaps if we didn't edit so many layers it would be better.