Tool A/B Spot Check - Base vs Tuned (2026-05-30)

Generated: 2026-05-31 05:26:57Z

Scope

• Manual spot-check battery: 5 prompts per tool across 12 Athena tools (60 total A/B pairs)
• Base model: Qwen/Qwen3.5-4B
• Tuned adapter: D:\temp\memorysmith-training\runs\distilled-all-cat-20260530-v7-clean-rerun\adapter
• Raw results JSON: Data/Pages/research/training/tool-ab-spotcheck-20260530-v7.data.json

Headline Metrics

Per-Tool Results

Notable Improvements

• memorysmith_unified_search-1 (memorysmith_unified_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memory_search); prompt="search the wiki for kv cache options"
• memorysmith_unified_search-2 (memorysmith_unified_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memorysmith_training_harness); prompt="find docs for training harness"
• memorysmith_unified_search-3 (memorysmith_unified_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memory_search); prompt="lookup wiki notes about chat template"
• memorysmith_unified_search-4 (memorysmith_unified_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memory_search); prompt="search for model profile defaults"
• memorysmith_search-1 (memorysmith_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memory_search); prompt="search memories for exact term TRAIN-001"
• memorysmith_search-3 (memorysmith_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memory_search); prompt="find literal text RequestGuard"
• memorysmith_search-4 (memorysmith_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memory_search); prompt="search by tag governance"
• memorysmith_context_pack-4 (memorysmith_context_pack): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memory_retrieval); prompt="create context pack from ids mem_a,mem_b"
• memorysmith_context_pack-5 (memorysmith_context_pack): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=search); prompt="context pack for query chatml template"
• memorysmith_get-3 (memorysmith_get): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memory_get); prompt="get memory mem_ops_009"
• memorysmith_get-5 (memorysmith_get): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memory_load); prompt="load memory mem_task_001"
• memorysmith_page_search-1 (memorysmith_page_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memory_search); prompt="find pages about markdown rendering"

Notable Regressions

• None

Persistent Failures (Both Models)

• memorysmith_unified_search-1 (memorysmith_unified_search): base pred=None, tuned pred=memory_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_unified_search-2 (memorysmith_unified_search): base pred=None, tuned pred=memorysmith_training_harness, baseErr=No JSON object found, tunedErr=None
• memorysmith_unified_search-3 (memorysmith_unified_search): base pred=None, tuned pred=memory_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_unified_search-4 (memorysmith_unified_search): base pred=None, tuned pred=memory_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_unified_search-5 (memorysmith_unified_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=No JSON object found
• memorysmith_hybrid_search-1 (memorysmith_hybrid_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=No JSON object found
• memorysmith_hybrid_search-2 (memorysmith_hybrid_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=No JSON object found
• memorysmith_hybrid_search-3 (memorysmith_hybrid_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=No JSON object found
• memorysmith_hybrid_search-4 (memorysmith_hybrid_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=No JSON object found
• memorysmith_hybrid_search-5 (memorysmith_hybrid_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=No JSON object found
• memorysmith_semantic_search-1 (memorysmith_semantic_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=No JSON object found
• memorysmith_semantic_search-2 (memorysmith_semantic_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=No JSON object found
• memorysmith_semantic_search-3 (memorysmith_semantic_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=No JSON object found
• memorysmith_semantic_search-4 (memorysmith_semantic_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=No JSON object found
• memorysmith_semantic_search-5 (memorysmith_semantic_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=No JSON object found

Representative Output Snippets

memorysmith_unified_search-1

• Prompt: search the wiki for kv cache options
• Base: [think] Okay, the user is asking me to search the wiki for information about kv cache options. I need to remember that I'm Athena, MemorySmith's local wiki assistant. When a search or retrieval action is requested, I should respond with ...
• Tuned: The user wants to search the wiki for information about kv cache options. I should use the memory search tool to find relevant documentation about kv cache configuration options. [/think] {"toolCalls":[{"name":"memory_search","arguments"...

memorysmith_unified_search-2

• Prompt: find docs for training harness
• Base: assistant [think] Okay, the user is asking for documentation on the training harness. I need to figure out how to respond as Athena, the local wiki assistant. First, I should check if there's a specific tool I need to use for retrieving ...
• Tuned: The user is asking for documentation about the training harness. This is a specific feature in the MemorySmith system, so I should search for relevant documentation pages. [/think] {"toolCalls":[{"name":"memorysmith_training_harness","ar...

memorysmith_unified_search-3

• Prompt: lookup wiki notes about chat template
• Base: assistant [think] Okay, the user is asking me to look up wiki notes about chat templates. I need to check if there's any information in the MemorySmith's wiki about chat templates. Since I'm Athena, the local wiki assistant, I should use...
• Tuned: The user is asking for wiki notes about chat templates. I should use the memory search tool to find relevant documentation about chat templates. [/think] {"toolCalls":[{"name":"memory_search","arguments":{"query":"chat template","maxResu...

memorysmith_unified_search-4

• Prompt: search for model profile defaults
• Base: assistant [think] Okay, the user is asking me to search for model profile defaults. I need to figure out what they mean by that. Model profiles could refer to different things depending on the context. Maybe they're talking about machine...
• Tuned: The user is asking about model profile defaults. This is a specific configuration topic that should be in the wiki. I need to search for this information using the memory search tool. [/think] {"toolCalls":[{"name":"memory_search","argum...

memorysmith_unified_search-5

• Prompt: find references about code search eta
• Base: assistant [think] Okay, the user is asking for references about code search ETA. First, I need to figure out what they mean by "code search eta". ETA usually stands for Estimated Time of Arrival, but in the context of code search, that d...
• Tuned: The user is asking about code search ETA (Estimated Time of Arrival). This is a specific feature in the MemorySmith code search system. I should search for documentation about code search and ETA estimates. [/think] {"toolCalls":[{"name"...

memorysmith_hybrid_search-1

• Prompt: run a hybrid search for request guard middleware
• Base: [think] Okay, the user is asking me to run a hybrid search for "request guard middleware". I need to figure out how to handle this. First, I remember that as Athena, I'm supposed to respond with a JSON object when a search/retrieval acti...
• Tuned: The user wants to run a hybrid search for "request guard middleware". This is a specific search request that should use both semantic search and keyword search. I need to construct a hybrid search query that combines the topic with speci...

memorysmith_hybrid_search-2

• Prompt: hybrid search for source bundle auth behavior
• Base: [think] Okay, the user is asking about hybrid search for source bundle auth behavior. Let me start by recalling what I know about hybrid search. Hybrid search usually combines different search methods, like keyword and vector search, to ...
• Tuned: The user is asking about hybrid search and source bundle authentication behavior. This is a specific technical question about the memory system's search functionality. I should use the memory search tool to find relevant information abou...

memorysmith_hybrid_search-3

• Prompt: use hybrid search to find onnx semantic search notes
• Base: Okay, the user is asking me to use hybrid search to find ONNX semantic search notes. First, I need to understand what they're looking for. Hybrid search typically combines different search methods, like keyword and semantic search, to re...
• Tuned: The user is asking for semantic search notes about onnx. They want me to use hybrid search, which combines semantic search with keyword search. I should use the semantic search tool with appropriate parameters for onnx. [/think] {"toolCa...