Why the result decode is all <unused42>? something wrong with the tokenizer?

#6
by AxolsWebAI - opened

[
{
"generated_text": [
{
"role": "user",
"content": "Generate a Python code snippet for the Bubble Sort algorithm, make the answer concise."
},
{
"role": "assistant",
"content": "<unused42><unused42>"
}
]
}
]

Sign up or log in to comment