Show HN: LLM Playground, compare results side by side

#1

Post by kwhinnery » Thu, Apr 20, 2023, 8:18 PM UTC

Show HN: LLM Playground, compare results side by side
retool.com

Re: Show HN: LLM Playground, compare results side by side

#2

Post by kwhinnery » Thu, Apr 20, 2023, 8:33 PM UTC

I was surprised by the quality of the results from Claude, I hadn't used it until recently. I asked it for strategic advice about playing a game of Magic: The Gathering and it gave what felt like pretty nuanced advice (talked about "resolving" spells, which is parlance only really used by avid players).

Re: Show HN: LLM Playground, compare results side by side

#3

Post by xnx » Thu, Apr 20, 2023, 9:55 PM UTC

Nice work! I'm equal parts impressed and terrified that CSS can do this.

Re: Show HN: LLM Playground, compare results side by side

#4

Post by cl42 » Thu, Apr 20, 2023, 10:50 PM UTC

kwhinnery wrote: post #2

I was surprised by the quality of the results from Claude, I hadn't used it until recently. I asked it for strategic advice about playing a game of Magic: The Gathering and it gave what felt like pretty nuanced advice (talked about "resolving" spells, which is parlance only really used by avid players).

Anthropic recently released a new version (v1.3) a few days ago. I've been using it for Python code generation and it very good.

Re: Show HN: LLM Playground, compare results side by side

#5

Post by season2episode3 » Fri, Apr 21, 2023, 12:31 AM UTC

Does this use your own API keys? Several of the models basically don't load for me, and I assume it's being hugged to death.