About augmentation with Change specific numbers.

when I use "Change specific numbers:" to augment problem using internlm2-math-20b to, the model answering the question instead of augment it~

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("internLLMATH", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("internLLMATH", trust_remote_code=True, torch_dtype=torch.float16).cuda()
model = model.eval()
response, history = model.chat(tokenizer, "Change specific numbers: Find the center of the circle with equation $x^2 - 6x + y^2 + 2y = 9$", history=[], meta_instruction="")
print(response)

We complete the square for both $x$ and $y$ terms. For the $x$ terms, we add $(6/2)^2 = 9$ to both sides. For the $y$ terms, we add $(2/2)^2 = 1$ to both sides. This gives us:
$(x^2 - 6x + 9) + (y^2 + 2y + 1) = 9 + 9 + 1$
$(x - 3)^2 + (y + 1)^2 = 19$
So, the center of the circle is $\boxed{(3, -1)}$.
The answer is: (3,-1).

Start by reproducing the behavior with the provided Python script and the model.chat call using the "Change specific numbers:" prompt. Inspect the model's augmentation behavior and determine what distinguishes augmentation from answering; done means this prompt changes the problem's numbers instead of returning a solved response.

索引モデルが issue の本文から書いたものです。