DebateLab at KIT

Enterprise
university
Activity Feed

AI & ML interests

computational philosophy, formal argumentation theory, normative models of reasoning

Recent Activity

DebateLabKIT's activity

ggbetzΒ 
posted an update about 1 month ago
view post
Post
1867
We've just released syncIALO -- a multi-purpose synthetic debate and argument mapping corpus with more than 600k arguments:

πŸ“ Blog article: https://huggingface.co/blog/ggbetz/introducing-syncialo
πŸ›’οΈ Dataset: DebateLabKIT/syncialo-raw
πŸ‘©β€πŸ’» Code: https://github.com/debatelab/syncIALO

πŸ€— Hugging Face has sponsored the syncIALO project through inference time / compute credits. πŸ™ We gratefully acknowledge the generous support. 🫢
ggbetzΒ 
posted an update 6 months ago
view post
Post
1499
Hi, just a brief follow-up on our Guided Reasoning (GuiR) system:

I've created a template space that facilitates testing:

1. Duplicate space logikon/guir-chat
2. Setup your own inference servers and provide details in config file
3. Add api keys as secrets
4. Your personal GuiR playground is ready

Cheers, Gregor
ggbetzΒ 
posted an update 6 months ago
view post
Post
1203
🧭 Guided Reasoning

πŸ‘‹Hi everyone,

We've been releasing Guided Reasoning:

Our AI guides walk your favorite LLM through complex reasoning problems.

🎯 Goals:

1️⃣ Reliability. AIs consistently follow reasoning methods.
2️⃣ Self-explainability. AIs see reasoning protocols and can explain internal deliberation.
3️⃣ Contestability. Users may amend AI reasoning and revise plausibility assessments.

Try out Guided Reasoning with our light demo chatbot, powered by πŸ€— HuggingFace's free Inference Api and small LLMs. (Sorry for poor latency and limited availability -- we are currently searching for πŸ’Έ compute sponsors to run more powerful models, faster, and optimize guided reasoning performance.)

Built on top of Logikon's open-source AI reasoning analytics.

Demo chat app: logikon/benjamin-chat
Github: https://github.com/logikon-ai/logikon
Technical report: https://arxiv.org/abs/2408.16331

➑️ Check it out and get involved! Looking forward to hearing from you.
ggbetzΒ 
posted an update 12 months ago
view post
Post
1446
πŸ₯‡Open CoT Leaderboard

We're delighted to announce the [Open CoT Leaderboard]( logikon/open_cot_leaderboard) on πŸ€— Spaces.

Unlike other LLM performance leaderboards, the Open CoT Leaderboard is not tracking absolute benchmark accuracies, but relative **accuracy gains** due to **chain-of-thought**.

Eval datasets that underpin the leaderboard are hosted [here](https://huggingface.co/cot-leaderboard).

Feedback and suggestions more than welcome.

@clefourrier
Β·