Spaces:
Running
Running
Commit History
-
feat: use llama-3.3-70b as primary model and clean prompt Running
24a3ca9 -
fix: inject {context} block into _SISTEM_PROMPT_TEMPLATE
3c293c7 -
fix: synthesis instructions in system prompt
9c8c152 -
fix: prompt synthesis rules for generic doctrinal legal queries
f2a8d1d -
feat: concept-based legal query understanding and soft domain prior
0adeaaf -
feat: implement canonical legal prompt, citation engine, and terminology normalizer
c6b1dfe
-
feat: integrate dynamic LLM legal query expansion & enable cross-encoder reranking
01e085b -
fix: timeout + scope cache + empty yanit fallback
e53f67b -
fix: max_tokens 2500 + empty-yanit fallback retry for reasoning models
075c127 -
fix: LLM scope check max_tokens 5->400 + regex EVET/HAYIR detection
6693699 -
feat: LLM-based scope detection replaces keyword-only filter
9c90930 -
fix: scope filter vergi/boşan/ceza + think-tag stripper
dd36569 -
fix: strengthen think-tag stripping for unclosed/stray tags
b335467 -
fix(generator): add model fallback and backoff retry for RateLimitError
425b318 -
fix(backend): groq model fallback and aksaklik hatasi cozumu
798537f
-
feat: Implement Hallmark Theme Engine, Multi-Tenant Client Config, Admin Settings, and White-Label Widget MVP
1c09698
-
feat: update chatbot UI to match poster requirements and backend retrieval improvements
516a4dc
-
feat: integrate real backend metrics into admin dashboard and add documents upload page
db42b4c -
feat: add dynamic PDF upload and site-document RAG capabilities
d0204d1
-
chore: production temizliği, gereksiz betiklerin silinmesi, yorum optimizasyonu ve README güncellemesi
e0ce37f -
initial clean commit
604ce70