Unverified50% confidenceBenchmarkExact time
The developer tested the gemma3:4b model (temperature 0.1) with 30 manually labeled items: under the v1 prose prompt configuration, brand accuracy was 97% with 1/4 fabricated brands; after adding a JSON Schema in v2, brand accuracy was 100% with 0/4 fabricated brands.
1
Sources
50%
Confidence
Long-term
Relevance
10/19/2026
First Seen
Sources
Related Entities
Related Claims
Unverifiedgemma-4-31b-it is an instruction-tuned version with multimodal capabilities from Google's Gemma series, with approximately 31 billion parameters.63% similarUnverifiedAfter adding a sampling configuration of temperature 0.6, top_p 0.95, top_k 20 to the serving JSON, the looping problem of the IQ1_M-quantized coding model under greedy decoding was fixed, and lines 20-24 of runconfig.py confirm that these keys are correctly received.62% similarUnverifiedThe test scoring used a weighted formula: 0.20×readability + 0.35×design execution and style match + 0.25×typography and completeness + 0.20×attractiveness, rounded to one decimal place.61% similarUnverifiedMuse Glimmer 30B scored 51.2 on the SWE benchmark, meaning it can solve approximately half of real-world software engineering problems.61% similarUnverifiedgemma-4-26b-a4b 模型名称中的 'a4b' 表示该模型采用了混合专家（MoE）架构，激活参数量约为4B60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1067469API
curl https://kongchang.com/api/v1/knowledge/claims/1067469MCP
get_claim(id=1067469)