Antoine/ATOCore
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

config: default LLM extractor model haiku -> sonnet

Browse Source 
Haiku was producing noisy candidates (31% accept rate on first
triage). Sonnet should give tighter extraction with fewer false
positives while still catching the same durable-fact patterns.
Override: ATOCORE_LLM_EXTRACTOR_MODEL=haiku to revert.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
Anto01
2026-04-12 07:31:34 -04:00
parent 146f2e4a5e
commit b0fde3ee60
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/atocore/memory/extractor_llm.py
View File

@@ -27,7 +27,7 @@ Configuration:
- Requires the ``claude`` CLI on PATH (``claude --version`` should work). - Requires the ``claude`` CLI on PATH (``claude --version`` should work).
- ``ATOCORE_LLM_EXTRACTOR_MODEL`` overrides the model alias (default - ``ATOCORE_LLM_EXTRACTOR_MODEL`` overrides the model alias (default
``haiku``). ``sonnet``).
- ``ATOCORE_LLM_EXTRACTOR_TIMEOUT_S`` overrides the per-call timeout - ``ATOCORE_LLM_EXTRACTOR_TIMEOUT_S`` overrides the per-call timeout
(default 90 seconds — first invocation is slow because Node.js (default 90 seconds — first invocation is slow because Node.js
startup plus OAuth check is non-trivial). startup plus OAuth check is non-trivial).
@@ -65,7 +65,7 @@ from atocore.observability.logger import get_logger
log = get_logger("extractor_llm") log = get_logger("extractor_llm")
LLM_EXTRACTOR_VERSION = "llm-0.2.0" LLM_EXTRACTOR_VERSION = "llm-0.2.0"
DEFAULT_MODEL = os.environ.get("ATOCORE_LLM_EXTRACTOR_MODEL", "haiku") DEFAULT_MODEL = os.environ.get("ATOCORE_LLM_EXTRACTOR_MODEL", "sonnet")
DEFAULT_TIMEOUT_S = float(os.environ.get("ATOCORE_LLM_EXTRACTOR_TIMEOUT_S", "90")) DEFAULT_TIMEOUT_S = float(os.environ.get("ATOCORE_LLM_EXTRACTOR_TIMEOUT_S", "90"))
MAX_RESPONSE_CHARS = 8000 MAX_RESPONSE_CHARS = 8000
MAX_PROMPT_CHARS = 2000 MAX_PROMPT_CHARS = 2000