fix: gpt-oss prompt preloading (#496)

Author: giladgd

src/evaluator/LlamaChatSession/LlamaChatSession.ts

@@ -854,6 +854,7 @@ export class LlamaChatSession {
     public async preloadPrompt(prompt: string, options: LLamaChatPreloadPromptOptions = {}): Promise<void> {
         await this.completePromptWithMeta(prompt, {
             ...options,
+            completeAsModel: false,
             maxTokens: 0
         });
     }
@@ -1009,7 +1010,9 @@ export class LlamaChatSession {
                             seed,
                             tokenBias,
                             customStopTriggers,
-                            maxTokens,
+                            maxTokens: maxTokens == null
+                                ? undefined
+                                : Math.min(1, maxTokens), // regular prompting ignores `maxTokens: 0`
                             temperature,
                             trimWhitespaceSuffix,
                             contextShift: {

templates/electron-typescript-react/package.json

@@ -13,7 +13,7 @@
     "_postinstall": "npm run models:pull",
     "models:pull": "node-llama-cpp pull --dir ./models \"{{modelUriOrUrl|escape|escape}}\"",
     "start": "vite dev",
-    "start:inspect": "ENABLE_INSPECT=true vite dev",
+    "start:inspect": "cross-env ENABLE_INSPECT=true vite dev",
     "start:build": "electron ./dist-electron",
     "prebuild": "rimraf ./dist ./dist-electron ./release",
     "build": "tsc && vite build && electron-builder --config ./electron-builder.ts",
@@ -42,6 +42,7 @@
     "@types/react-dom": "^19.1.3",
     "@types/semver": "^7.7.0",
     "@vitejs/plugin-react": "^4.4.1",
+    "cross-env": "^10.0.0",
     "electron": "^36.2.0",
     "electron-builder": "^26.0.12",
     "eslint": "^9.26.0",