nanochat/modal
Manmohan Sharma 3c5a815f92
fix: inject Tavily snippet text into grounding suffix
Previous suffix 'Based on the search results above, ' still let the model write meta-commentary ('1 person provided an answer') instead of content. Now we extract the first result's snippet from the tool output and seed the suffix with 'Here's what I found: {snippet}' — the model's next tokens continue from real prose rather than reasoning about the JSON structure.
2026-04-22 16:12:59 -07:00
..
_model.py feat: deploy d24 SFT + polished UI redesign with dark mode (#39) 2026-04-16 19:55:16 -04:00
_query_classifier.py fix: veto matches shorthand 'u' and 'r' for you/are 2026-04-22 16:10:59 -07:00
_tokenizer.py feat: deploy d24 SFT + polished UI redesign with dark mode (#39) 2026-04-16 19:55:16 -04:00
_tools.py fix(tools): enable Tavily include_answer and fix UI overflow 2026-04-22 14:20:47 -07:00
serve.py fix: inject Tavily snippet text into grounding suffix 2026-04-22 16:12:59 -07:00