diff --git a/src/llama.cpp b/src/llama.cpp
index 0a34f46e6..4c6c9678e 100644
--- a/src/llama.cpp
+++ b/src/llama.cpp
@@ -3994,7 +3994,9 @@ struct llama_model_loader {
                 throw std::runtime_error(format("%s is not a float32, int32 array", key.c_str()));
         }
 
-        GGML_ASSERT(arr_info.length <= N_MAX);
+        if (arr_info.length > N_MAX) {
+            throw std::runtime_error(format("array length %u for key %s exceeds max %u", (uint32_t) arr_info.length, key.c_str(), (uint32_t) N_MAX));
+        }
 
         std::copy((const T*)arr_info.data, (const T *)arr_info.data + arr_info.length, result.begin());
 
@@ -4030,8 +4032,6 @@ struct llama_model_loader {
     // get array of n <= N_MAX elements, or a single element repeated n times
     template<typename T, size_t N_MAX>
     bool get_key_or_arr(const std::string & key, std::array<T, N_MAX> & result, uint32_t n, const bool required = true) {
-        GGML_ASSERT(n <= N_MAX);
-
         const int kid = gguf_find_key(meta, key.c_str());
 
         if (kid < 0) {
@@ -4041,6 +4041,10 @@ struct llama_model_loader {
             return false;
         }
 
+        if (n > N_MAX) {
+            throw std::runtime_error(format("n > N_MAX: %u > %u for key %s", (uint32_t) n, (uint32_t) N_MAX, key.c_str()));
+        }
+
         if (gguf_get_kv_type(meta, kid) == GGUF_TYPE_ARRAY) {
             struct GGUFMeta::ArrayInfo arr_info =
                 GGUFMeta::GKV<GGUFMeta::ArrayInfo>::get_kv(meta, kid);