Llama 4 was a giant disappointment, let's wait for Qwen 3.
Burnt out and unimpressed, anyone else?
DeepSeek V3 0324 is so goddamn horny.
@36, doing video calls today.
Deepseek-R1 prompt for ERP
Deepseek-R1 is the filthiest model I’ve ever used
The first time I've felt a LLM wrote *well*, not just well *for a LLM*.
@ 0:14
How can U.S states like Florida enforce ID verification laws against non U.S websites like Pormhub? What is the *mechanism* of enforcement?
How can individual states like Florida force sites like Pornhub to require age verification?
@0:10 buttplug cum
EXL2 Inference Quality Issues
What format are you using for your RP/ERP character descriptions?
@ 0:35
Magnum v3 - 9b (gemma and chatml)
Choker
Llama 3.1 Discussion and Questions Megathread
L3-70B-Euryale-v2.1 - A Finetune for Roleplay / Creative Purposes
Bliss
Moist Miqu 70B - the largest, moistiest model yet from BeaverAI! From the creators of CreamPhi, Moistral, Coomand-R, and Llama 3Some!
Deepdive into Llama 3’s arena wins against bigger models: “As prompts get challenging, the gap between Llama 3 against top-tier models becomes larger.”
Llama 3 8B instruct with fixed BPE tokenizer uploaded
Result: Llama 3 EXL2 quant quality compared to GGUF and Llama 2
Most Viewed Video of the Month
I think Command R+ is really underrated.
