Tool A/B Spot Check - 20260601-step01-v6-chatml-batched-t256 (Base vs Tuned)

Generated: 2026-06-01 07:17:14Z

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\v6-chatml-batched-20260601-step1\adapter
• Raw results JSON: Data/Pages/research/training/tool-ab-spotcheck-20260601-step01-v6-chatml-batched-t256.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=memorysmith_search); prompt="search the wiki for kv cache options"
• memorysmith_unified_search-3 (memorysmith_unified_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memorysmith_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=memorysmith_search); prompt="search for model profile defaults"
• memorysmith_unified_search-5 (memorysmith_unified_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memorysmith_search); prompt="find references about code search eta"
• memorysmith_hybrid_search-2 (memorysmith_hybrid_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memorysmith_search); prompt="hybrid search for source bundle auth behavior"
• memorysmith_hybrid_search-3 (memorysmith_hybrid_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memorysmith_search); prompt="use hybrid search to find onnx semantic search notes"
• memorysmith_hybrid_search-4 (memorysmith_hybrid_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memorysmith_search); prompt="hybrid search for maintenance proposal workflow"
• memorysmith_hybrid_search-5 (memorysmith_hybrid_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memorysmith_search); prompt="run hybrid search for task assignee update invariant"
• memorysmith_semantic_search-1 (memorysmith_semantic_search): base(match=False, env=False, pred=None) -> tuned(match=True, env=True, pred=memorysmith_semantic_search); prompt="semantic search for retrieval warning propagation"
• memorysmith_semantic_search-2 (memorysmith_semantic_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memorysmith_search); prompt="semantic search for pages lock navigation regression"
• memorysmith_semantic_search-3 (memorysmith_semantic_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=semantic_search); prompt="semantic search for sidebar collapse standard"
• memorysmith_semantic_search-4 (memorysmith_semantic_search): base(match=False, env=False, pred=None) -> tuned(match=False, env=True, pred=memorysmith_search); prompt="semantic search for source-link default app open"

Notable Regressions

• None

Persistent Failures (Both Models)

• memorysmith_unified_search-1 (memorysmith_unified_search): base pred=None, tuned pred=memorysmith_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_unified_search-2 (memorysmith_unified_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=toolCalls missing or empty
• memorysmith_unified_search-3 (memorysmith_unified_search): base pred=None, tuned pred=memorysmith_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_unified_search-4 (memorysmith_unified_search): base pred=None, tuned pred=memorysmith_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_unified_search-5 (memorysmith_unified_search): base pred=None, tuned pred=memorysmith_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_hybrid_search-1 (memorysmith_hybrid_search): base pred=None, tuned pred=None, baseErr=No JSON object found, tunedErr=JSON parse error: Expecting property name enclosed in double quotes: line 1 column 42 (char 41)
• memorysmith_hybrid_search-2 (memorysmith_hybrid_search): base pred=None, tuned pred=memorysmith_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_hybrid_search-3 (memorysmith_hybrid_search): base pred=None, tuned pred=memorysmith_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_hybrid_search-4 (memorysmith_hybrid_search): base pred=None, tuned pred=memorysmith_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_hybrid_search-5 (memorysmith_hybrid_search): base pred=None, tuned pred=memorysmith_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_semantic_search-2 (memorysmith_semantic_search): base pred=None, tuned pred=memorysmith_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_semantic_search-3 (memorysmith_semantic_search): base pred=None, tuned pred=semantic_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_semantic_search-4 (memorysmith_semantic_search): base pred=None, tuned pred=memorysmith_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_semantic_search-5 (memorysmith_semantic_search): base pred=None, tuned pred=semantic_search, baseErr=No JSON object found, tunedErr=None
• memorysmith_search-1 (memorysmith_search): base pred=None, tuned pred=memorysmith_search_memories, baseErr=No JSON object found, tunedErr=None

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 is asking me to search the wiki for information about "kv cache options". I need to use the memorysmith_search tool to search for this information. I should format the response as a single JSON object with toolCalls containing t...

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 a "training harness". This sounds like they want to search for or retrieve documentation related to a training harness feature. I should use the memorysmith_search tool to find relevant document...

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 me to look up wiki notes about "chat template". This is a search/retrieval action request, so I need to respond with exactly one JSON object containing a tool call for the memorysmith_search tool. I should format this ...

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 me to search for "model profile defaults". This appears to be a request for memorysmith search functionality. I need to respond with exactly one JSON object containing a tool call for the memorysmith search tool. Let m...

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 me to find references about "code search eta". This seems to be a request for information about a feature or tool called "code search eta" - likely related to memorysmith or Athena's capabilities. I should respond with...

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 is asking me to run a hybrid search for "request guard middleware". This is a search/retrieval action request, so I need to respond with exactly one JSON object containing a tool call for the search tool. I should format this as...

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 for source bundle auth behavior. This sounds like they want to search for information about authentication behavior related to source bundles in a hybrid search context. This is a memorysmith search...

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 me to use hybrid search to find onnx semantic search notes. This is a search/retrieval action request, so I need to respond with exactly one JSON object containing a tool call for the memorysmith_search tool. I should ...