Which Tokens Should SFT Actually Learn? A Token-Trimming Perspective on Mathematical Reasoning
arXiv:2609.09707v1 Announce Type: new Abstract: Supervised fine-tuning (SFT) applies a uniform cross-entropy loss to all target tokens, even though different tokens provide unequal learning signals for mathematical reasoning. This uniform treatment can over-sharpen already…
- 01.arXiv:2609.09707v1 Announce Type: new Abstract: Supervised fine-tuning (SFT) applies a uniform cross-entropy loss to all target tokens, even though different tokens provide unequal learning signals for mathematical reasoning
- 02.This uniform treatment can over-sharpen already…
This material is provided as an extended analytical summary.
Read full article at sourceRelated Intelligence
All stories →Anthropic opens its books for Wall Street, showing record revenue and mounting obligations
Anthropic is the first major AI lab to file an S-1 with the SEC, planning to list on the Nasdaq under the ticker "ANT." The numbers are strong, with $11.6 billion in quarterly revenue, a first-ever operating profit, and existing customers spending nine times more year over year.…
Создание агентной системы извлечения на базе Amazon Bedrock с CDK
AWS Machine Learning описывает создание агентной системы извлечения на базе Amazon Bedrock с использованием AWS CDK.
Кэширование запросов для ИИ-агентов на Amazon Bedrock
Кэширование запросов позволяет ИИ-агентам на Amazon Bedrock быть более эффективными.