Menu

#Preference

5 posts

Feed·
5 of 5 posts
When Generic Benchmarks Fail: Building a Sales-Domain Evaluation Bench from Scratch
🖼️
0

When Generic Benchmarks Fail: Building a Sales-Domain Evaluation Bench from Scratch

DEV Community·Nati A·about 1 month ago
#pZ6djcTx

How I built Tenacious-Bench — a 240-task domain-specific benchmark for a B2B sales agent — trained a SimPO LoRA judge, and lifted held-out preference accuracy from 14.9% (deterministic baseline) to 91.5% on the same 47-task slice.

15s
Read More
When Your Training Loss Is Lying to You Building a Tenacious-Specific Sales Outreach Benchmark Eyoel Nebiyu · May 2026
🖼️
0

When Your Training Loss Is Lying to You Building a Tenacious-Specific Sales Outreach Benchmark Eyoel Nebiyu · May 2026

DEV Community·Eyoel Nebiyu·about 1 month ago
#k271xe5Q
#agents#ai#llm#machinelearning#model#training

From Dev RSS Feed: When Your Training Loss Is Lying to You Building a Tenacious-Specific Sales Outreach Benchmark Eyoel Nebiyu · May 2026

15s
Read More
📰
0

Absolute paths for assets in iOS & Android Cordova apps

DEV Community: cordova·Tyler Smith·about 1 month ago
#dTmsTNKi
#dev#preference#l222#issuecomment#class#code

There's an abundance of outdated information that states you can only use relative paths for assets in Cordova mobile apps. That information is incorrect: you can use absolute paths for assets in Cordova by configuring the scheme and hostname within…

15s
Read More