5 ms·
Obviously it's not going to be of a paid tier 2T sized SOTA model quality, but it can probably roughly match Haiku at the very least. And for tasks that aren't
by moffkalast 8mo ago
Obviously it's not going to be of a paid tier 2T sized SOTA model quality, but it can probably roughly match Haiku at the very least. And for tasks that aren't super complex that's already enough.
Personally though, I find Qwen useless for anything but coding tasks because if its insufferable sycophancy. It's like 4o dialed up to 20, every reply starts with "You are absolutely right" with zero self awareness. And for coding, only the best model available is usually sensible to use otherwise it's just wasted time.
- Anduia 8mo agoThat's why I start any prompt to Qwen 3.5 with: persona: brief rude senior
- lemonginger 8mo agogamechanger
- amelius 8mo agoI'm using: persona: drunken sailor Because then at least the tone matches the quality of the output and I'm reminded of what I can expect.
- dlcarrier 8mo agoDoes it tend to break out into sea shanties?
- moffkalast 8mo agoBut then what do you do with it early in the morning?
- amelius 8mo agoFor starters, shave his belly with a rusty razor, obviously ;)
- em500 8mo agoThis also works persona: emotionless vulcan
- dlcarrier 8mo agoDoes "persona: air traffic controller" work? If I could set up a voice assistant that actually verifies commands, instead of assuming it heard everything correctly 100% of the time, it might even be useful.
- varispeed 8mo agopersona: party delegate in a rural province who doesn't want to be there
- 9wzYQbTYsAIc 8mo agopersona: fair witness https://fairwitness.bot/ https://fairwitness.bot/
- Chris2048 8mo agoYou just paste in that YAML? Is this an official llm config format that is parsed out?
- 9wzYQbTYsAIc 8mo agoYeah, just paste it in there - the LLM will figure it out. Play with it if you want to tweak the formatting - you could try JSON instead, but for readability I went with YAML.
- ranger_danger 8mo agowow I had no idea you could do that. this changes everything for me.
- ggregoire 8mo ago> I find Qwen useless for anything but coding tasks because if its insufferable sycophancy We use Qwen at work since 2.0 for text/image/video analysis (summarization, categorization, NER, etc), I think it's impressive. We ask for JSON and always ask "do not explain your response".
- andai 8mo ago>for coding, only the best model available is usually sensible to use otherwise it's just wasted time. I had the opposite experience. Gave a small model and a big model the same 3 tasks. Small model was done in 30 sec. Large model took 90 sec 3x longer and cost 3x more. Depending on the task, the benchies just tell you how much you are over-paying and over-waiting.
- wehadit 8mo agoIf you use the models like we execute coding tasks, older models outperform latest models. There's this prep tax that happens even before we start coding, i.e., extract requirements from tools, context from code, comments and decisions from conversations, ACs from Jira/Notion, stitch them together, design tailored coding standards and then code. If you automate the prep tax, the generated code is close to production ready code and may require 1-2 iterations max. I gave it a try and compared the results and found the output to be 92% accurate while same done on Claude Code gave 68% accuracy. Prep tax is the cue here
- itsTyrion 8mo agooh? I used it in t3 chat before, with traits `concise` `avoid unnecessary flattery/affirmation/praise` `witty` `feel free to match potential user's sarcasm` and it does use that sarcasm permission at times (I still dislike the way it generally communicates)