The model has been talking nonsense all along

[Deng-Xian-Sheng]

No description provided.

[ridgerchu]

Hi, Sorry for hear that, SpikeGPT is primarily an experimental product and has not undergone processes like Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF). These processes are essential for ensuring more natural and coherent conversations. Without them, SpikeGPT's ability to engage in natural chat is limited...

[fuxuelinwudi]

this model can't learn how to talk. it's decoding ability is like a shit. it's also what i want to say to rwkv.
this ark don't have a enough result for few-shot task, it needs lots of dataset to fit, so that it will get some good performance.
fucking rwkv, it don't either have half of performance of transformer, even than rnn!!!