Skip to content
Navigation Menu
Sign in
Appearance settings
Platform
AI CODE CREATION
GitHub Copilot
Write better code with AI
GitHub Copilot app
Direct agents from issue to merge
MCP Registry
Integrate external tools
DEVELOPER WORKFLOWS
Actions
Automate any workflow
Codespaces
Instant dev environments
Issues
Plan and track work
Code Review
Manage code changes
Code Quality
Enforce quality at merge
APPLICATION SECURITY
GitHub Advanced Security
Find and fix vulnerabilities
Code security
Secure your code as you build
Secret protection
Stop leaks before they start
EXPLORE
Why GitHub
Documentation
Blog
Changelog
Marketplace
View all features
Solutions
BY COMPANY SIZE
Enterprises
Small and medium teams
Startups
Nonprofits
BY USE CASE
App Modernization
DevSecOps
DevOps
CI/CD
View all use cases
BY INDUSTRY
Healthcare
Financial services
Manufacturing
Government
View all industries
View all solutions
Resources
EXPLORE BY TOPIC
AI
Software Development
DevOps
Security
View all topics
EXPLORE BY TYPE
Customer stories
Events & webinars
Ebooks & reports
Business insights
GitHub Skills
SUPPORT & SERVICES
Documentation
Customer support
Community forum
Trust center
Partners
View all resources
Open Source
COMMUNITY
GitHub Sponsors
Fund open source developers
PROGRAMS
Security Lab
Maintainer Community
GitHub Stars
Archive Program
REPOSITORIES
Topics
Trending
Collections
Enterprise
ENTERPRISE SOLUTIONS
Enterprise platform
AI-powered developer platform
AVAILABLE ADD-ONS
GitHub Advanced Security
Enterprise-grade security features
Copilot for Business
Enterprise-grade AI features
Premium Support
Enterprise-grade 24/7 support
Pricing
Search
/
Sign in
Sign up
Appearance settings
You signed in with another tab or window.
Reload
to refresh your session.
You signed out in another tab or window.
Reload
to refresh your session.
You switched accounts on another tab or window.
Reload
to refresh your session.
Dismiss alert
{{ message }}
Uh oh!
There was an error while loading.
Please reload this page
.
RouteWorks
/
RouterArena
Public
Notifications
You must be signed in to change notification settings
Fork
45
Star
126
Code
Issues
13
Pull requests
10
Actions
Projects
Security and quality
0
Insights
Additional navigation options
Code
Issues
Pull requests
Actions
Projects
Security and quality
Insights
Add cruq-router: self-consistency cascade router
- #197
#197
Closed
nabaruns
wants to merge 22 commits into
RouteWorks:main
RouteWorks/RouterArena:main
from
cruq-ai:cruq-router
cruq-ai/RouterArena:cruq-router
Copy head branch name to clipboard
Conversation
Commits
22
(22)
Checks
Files changed
Closed
Add cruq-router: self-consistency cascade router
#197
nabaruns
wants to merge 22 commits into
RouteWorks:main
RouteWorks/RouterArena:main
from
cruq-ai:cruq-router
cruq-ai/RouterArena:cruq-router
Copy head branch name to clipboard
Commits
Commits on Aug 20, 2026
Add cruq router (Phase 1): difficulty-tiered cheapest-sufficient policy
Show description for 5dc9f1b
nabaruns
committed
5dc9f1b
View commit details
Copy full SHA for 5dc9f1b
Browse repository at this point
Add Phase 2 pipeline: learned cheapest-sufficient capability predictor
Show description for cc9fba5
nabaruns
committed
cc9fba5
View commit details
Copy full SHA for cc9fba5
Browse repository at this point
Commits on Aug 21, 2026
Phase 2 hard-corpus iteration: harder sources, parallel labeling, grader fixes
Show description for c35106b
nabaruns
committed
c35106b
View commit details
Copy full SHA for c35106b
Browse repository at this point
Commits on Aug 22, 2026
Park RouterArena work: findings write-up + strong-tier scaffolding
Show description for 3f5bca3
nabaruns
committed
3f5bca3
View commit details
Copy full SHA for 3f5bca3
Browse repository at this point
Add parked findings write-up (phase2/FINDINGS.md)
nabaruns
committed
45ce914
View commit details
Copy full SHA for 45ce914
Browse repository at this point
Refute parked conclusion: cheap-model diversity, not a strong tier, lifts the oracle
Show description for d43f2be
nabaruns
committed
d43f2be
View commit details
Copy full SHA for d43f2be
Browse repository at this point
Commits on Aug 23, 2026
5-model router experiment: oracle lifts, selector cannot capture it
Show description for e8ba454
nabaruns
committed
e8ba454
View commit details
Copy full SHA for e8ba454
Browse repository at this point
Selector lever (a): calibration + per-model tau — marginal, collapses to best model
Show description for 0a52dd0
nabaruns
committed
0a52dd0
View commit details
Copy full SHA for 0a52dd0
Browse repository at this point
Domain-aware routing: best arena-S (0.708), capped by domain coverage
Show description for d8e04c1
nabaruns
committed
d8e04c1
View commit details
Copy full SHA for d8e04c1
Browse repository at this point
Official evaluator on k8s + critical validation: local scoring is unreliable
Show description for 936cf84
nabaruns
committed
936cf84
View commit details
Copy full SHA for 936cf84
Browse repository at this point
Fix local eval harness (missing config/eval_config) + corrected true numbers
Show description for 64b53d4
nabaruns
committed
64b53d4
View commit details
Copy full SHA for 64b53d4
Browse repository at this point
Commits on Aug 24, 2026
Domain router scored on fixed v3 official metric: 0.737 arena-S
Show description for 1135999
nabaruns
committed
1135999
View commit details
Copy full SHA for 1135999
Browse repository at this point
Widen corpus to 23 domains: regressed (0.737->0.733) — transfer, not coverage, is the cap
Show description for e7b30b5
nabaruns
committed
e7b30b5
View commit details
Copy full SHA for e7b30b5
Browse repository at this point
Online-probe cascade router: first to beat best-single accuracy (0.755)
Show description for 987b2d7
nabaruns
committed
987b2d7
View commit details
Copy full SHA for 987b2d7
Browse repository at this point
Self-consistency probe: best router (acc 0.760, honest arena-S 0.746)
Show description for cf70e5c
nabaruns
committed
cf70e5c
View commit details
Copy full SHA for cf70e5c
Browse repository at this point
Commits on Aug 25, 2026
Lower K cuts probe cost: K=4 is the new best (arena-S 0.7474)
Show description for 17a563d
nabaruns
committed
17a563d
View commit details
Copy full SHA for 17a563d
Browse repository at this point
Stronger escalation model: +1.2pt accuracy but lower arena-S (cost dominates)
Show description for 4f1ddbd
nabaruns
committed
4f1ddbd
View commit details
Copy full SHA for 4f1ddbd
Browse repository at this point
Commits on Aug 27, 2026
Add cruq-router: self-consistency cascade submission (full split)
Show description for 512401d
nabaruns
committed
512401d
View commit details
Copy full SHA for 512401d
Browse repository at this point
Add cruq-router robustness file; drop experimental prediction files
Show description for e8991c9
nabaruns
committed
e8991c9
View commit details
Copy full SHA for e8991c9
Browse repository at this point
Fix 2 empty free-form answers that failed submission validation
Show description for bcfd324
nabaruns
committed
bcfd324
View commit details
Copy full SHA for bcfd324
Browse repository at this point
Commits on Aug 28, 2026
Trim PR to the submission: config, predictions, router class
Show description for 5f7c09e
nabaruns
committed
5f7c09e
View commit details
Copy full SHA for 5f7c09e
Browse repository at this point
Normalize model_cost.json trailing newline (end-of-file-fixer)
nabaruns
committed
a30e400
View commit details
Copy full SHA for a30e400
Browse repository at this point
You can’t perform that action at this time.
