Skip to content
Open
Show file tree
Hide file tree
Changes from 12 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
The table of contents is too big for display.
Diff view
Diff view
  •  
  •  
  •  
The diff you're trying to view is too large. We only load the first 3000 changed files.
33 changes: 33 additions & 0 deletions .github/ISSUE_TEMPLATE/benchmark-run.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,33 @@
---
name: "Benchmark run (OpenRouter)"
description: "Request a benchmark run for a single OpenRouter model"
title: "Benchmark: <author>/<model> (OpenRouter)"
body:
- type: markdown
attributes:
value: |
Submit a benchmark request for a single OpenRouter model.

- Only org members/collaborators can trigger runs.
- A workflow will validate this issue when the `run-benchmark` label is applied.

- type: textarea
id: models
attributes:
label: Model (OpenRouter URL or slug)
description: |
Exactly one model. Accepted formats:
- https://openrouter.ai/<author>/<model>
- <author>/<model>
placeholder: |
https://openrouter.ai/mistralai/mistral-large-2512
validations:
required: true

- type: textarea
id: notes
attributes:
label: Notes (optional)
description: Any context for reviewers / maintainers.
validations:
required: false
15 changes: 15 additions & 0 deletions datasets/assist-de/dataset_card.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
---
name: assist-de
language: de
version: v2
description: |-
Ein Datensatz zum Testen der Home Assistant LLM API. Die Häuser in diesem
Datensatz wurden synthetisch mit gpt-3.5 generiert und anschließend manuell
kuratiert, um die Home Assistant Intents zur Gerätesteuerung zu testen. Die
Sätze wurden absichtlich schwieriger gestaltet als die bestehende
Assistenten-NLP, um die Denkfähigkeiten größerer Modelle zu demonstrieren.
urls:
- https://github.com/allenporter/home-assistant-datasets/tree/main/datasets/assist-de
- https://developers.home-assistant.io/blog/2024/05/20/llm-api/
# Erhöhe die Standard-Anzahl der Durchläufe für ein niedrigeres Konfidenzintervall
count: 5
Loading
Loading