Fix gating: per-element to match g_proj weights (was True/per-head) (#1)

- Fix gating: per-element to match g_proj weights (was True/per-head) (00ec45ba7f76f37003190cc025442957f7e2164c)

config.json

@@ -39,7 +39,7 @@
   "tie_word_embeddings": false,
   "use_cache": true,
   "torch_dtype": "bfloat16",
-  "gating": true,
+  "gating": "per-element",
   "sliding_window": 0,
   "rope_parameters": {
     "full_attention": {