Show HN: Perplexity (llama3 70B) Inline Bot on Telegram
telegrambots.ai
Show HN: Perplexity (llama3 70B) Inline Bot on Telegram
1–10 of 13 posts
Re: Show HN: Perplexity (llama3 70B) Inline Bot on Telegram
#2Word of caution when using perplexity to summarize URLs. I have tried the API and had lots of hallucinations.
It worked better in the UI, I only had it with the API.
Re: Show HN: Perplexity (llama3 70B) Inline Bot on Telegram
#3Is this Perplexity the search engine? Why does it mention Llama in the title?
Re: Show HN: Perplexity (llama3 70B) Inline Bot on Telegram
#4Is this Perplexity the search engine? Why does it mention Llama in the title?
I'm guessing because perplexity is now based on llama 3. If you look at their labs [1] the choices of models seems to suggest this.
Re: Show HN: Perplexity (llama3 70B) Inline Bot on Telegram
#5Is this Perplexity the search engine? Why does it mention Llama in the title?
I'm guessing because perplexity is now based on llama 3. If you look at their labs [1] the choices of models seems to suggest this. [1] https://labs.perplexity.ai
Sure, but it's a bit too "under the hood" to mention in something like this, no?
Re: Show HN: Perplexity (llama3 70B) Inline Bot on Telegram
#6Word of caution when using perplexity to summarize URLs. I have tried the API and had lots of hallucinations. It worked better in the UI, I only had it with the API.
Yeah, I don’t know why their API is inferior to the UI. Pretty disappointing. I’ve had better luck with my own “clone” that makes a web search and summarizes the results into an answer.
Re: Show HN: Perplexity (llama3 70B) Inline Bot on Telegram
#7Earlier quoted context omitted.
I'm guessing because perplexity is now based on llama 3. If you look at their labs [1] the choices of models seems to suggest this. [1] https://labs.perplexity.ai
Sure, but it's a bit too "under the hood" to mention in something like this, no?
I like under the hood info, that is why I am in hackernews :)
Re: Show HN: Perplexity (llama3 70B) Inline Bot on Telegram
#8Word of caution when using perplexity to summarize URLs. I have tried the API and had lots of hallucinations. It worked better in the UI, I only had it with the API.
I've been evaluating OpenAI (GPT-3.5 and -4o) for a project at my company and I have seen differences in quality between the api and the web version as well. Same prompts. The api version seems to vary a lot more in quality.
Re: Show HN: Perplexity (llama3 70B) Inline Bot on Telegram
#9Is this Perplexity the search engine? Why does it mention Llama in the title?
The money is in the app layer and compute. There's no reason to train foundation models to compete with Anthropic, OpenAI, Meta ect ect. Its the equivalent of lighting money on fire.
Re: Show HN: Perplexity (llama3 70B) Inline Bot on Telegram
#10Word of caution when using perplexity to summarize URLs. I have tried the API and had lots of hallucinations. It worked better in the UI, I only had it with the API.
I've been evaluating OpenAI (GPT-3.5 and -4o) for a project at my company and I have seen differences in quality between the api and the web version as well. Same prompts. The api version seems to vary a lot more in quality.
Play around with temperature. Summarization is quite a defined task, so lower the temp a lot.