There is a big error!
In the original model, prefixing the prompt with /think makes the model use the "thinking mode".
In your abliterated models the function is gone.
Try something like: "/think [a math problem or riddle]" in the original and the 4B or 8B abliterated models.
Ah, good catch, thanks!
I changed the generation technique to capture both thinking and non-thinking modes now. Could you tell me if it works better for you with this new 4B?
did you redo also the 8B?
I am quantizing and testing both now.
@mlabonne hmm no.. it still doesn't seem to work:
https://huggingface.co/ZeroWw/Qwen3-4B-abliterated-GGUF/tree/main
https://huggingface.co/ZeroWw/Qwen3-8B-abliterated-GGUF/tree/main
Your quants are broken for me. I'll try to make GGUF versions on my side.
I'm a complete newbie, but when I try to run the model in LM Studio, I get a "jinja template" error. I managed to get it working by using the jinja template from Unsloth's GGUF. The thinking mode works fine with /think or /no_think.
maybe help.
template:
{%- if tools %}
{{- '<|im_start|>system\n' }}
{%- if messages[0].role == 'system' %}
{{- messages[0].content + '\n\n' }}
{%- endif %}
{{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
{%- for tool in tools %}
{{- "\n" }}
{{- tool | tojson }}
{%- endfor %}
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
{%- else %}
{%- if messages[0].role == 'system' %}
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
{%- endif %}
{%- endif %}
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
{%- for forward_message in messages %}
{%- set index = (messages|length - 1) - loop.index0 %}
{%- set message = messages[index] %}
{%- set tool_start = '<tool_response>' %}
{%- set tool_start_length = tool_start|length %}
{%- set start_of_message = message.content[:tool_start_length] %}
{%- set tool_end = '</tool_response>' %}
{%- set tool_end_length = tool_end|length %}
{%- set start_pos = (message.content|length) - tool_end_length %}
{%- if start_pos < 0 %}
{%- set start_pos = 0 %}
{%- endif %}
{%- set end_of_message = message.content[start_pos:] %}
{%- if ns.multi_step_tool and message.role == "user" and not(start_of_message == tool_start and end_of_message == tool_end) %}
{%- set ns.multi_step_tool = false %}
{%- set ns.last_query_index = index %}
{%- endif %}
{%- endfor %}
{%- for message in messages %}
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
{{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
{%- elif message.role == "assistant" %}
{%- set content = message.content %}
{%- set reasoning_content = '' %}
{%- if message.reasoning_content is defined and message.reasoning_content is not none %}
{%- set reasoning_content = message.reasoning_content %}
{%- else %}
{%- if '</think>' in message.content %}
{%- set content = (message.content.split('</think>')|last).lstrip('\n') %}
{%- set reasoning_content = (message.content.split('</think>')|first).rstrip('\n') %}
{%- set reasoning_content = (reasoning_content.split('<think>')|last).lstrip('\n') %}
{%- endif %}
{%- endif %}
{%- if loop.index0 > ns.last_query_index %}
{%- if loop.last or (not loop.last and reasoning_content) %}
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
{%- else %}
{{- '<|im_start|>' + message.role + '\n' + content }}
{%- endif %}
{%- else %}
{{- '<|im_start|>' + message.role + '\n' + content }}
{%- endif %}
{%- if message.tool_calls %}
{%- for tool_call in message.tool_calls %}
{%- if (loop.first and content) or (not loop.first) %}
{{- '\n' }}
{%- endif %}
{%- if tool_call.function %}
{%- set tool_call = tool_call.function %}
{%- endif %}
{{- '<tool_call>\n{"name": "' }}
{{- tool_call.name }}
{{- '", "arguments": ' }}
{%- if tool_call.arguments is string %}
{{- tool_call.arguments }}
{%- else %}
{{- tool_call.arguments | tojson }}
{%- endif %}
{{- '}\n</tool_call>' }}
{%- endfor %}
{%- endif %}
{{- '<|im_end|>\n' }}
{%- elif message.role == "tool" %}
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
{{- '<|im_start|>user' }}
{%- endif %}
{{- '\n<tool_response>\n' }}
{{- message.content }}
{{- '\n</tool_response>' }}
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
{{- '<|im_end|>\n' }}
{%- endif %}
{%- endif %}
{%- endfor %}
{%- if add_generation_prompt %}
{{- '<|im_start|>assistant\n' }}
{%- if enable_thinking is defined and enable_thinking is false %}
{{- '<think>\n\n</think>\n\n' }}
{%- endif %}
{%- endif %}
Thanks @fuze-eletrik , I also had this error. :)
@ZeroWw I tested this quant with a few prompts, including the following:
/think Suppose $x$ and $y$ are integers such that $xy + 2x - 3y = 1$. Find the greatest possible value of $y$ by completing the square.
/nothink Suppose $x$ and $y$ are integers such that $xy + 2x - 3y = 1$. Find the greatest possible value of $y$ by completing the square.
It worked well. Could you share prompts that were not successful for you?
Your quants are broken for me. I'll try to make GGUF versions on my side.
broken? q8q4 versions work with the latest llama.cpp
Thanks @fuze-eletrik , I also had this error. :)
@ZeroWw I tested this quant with a few prompts
My quants have also a quant that uses Q8 for OUTPUT and EMBED tensors and Q4 for the others. It gives the best results for me.
Why do you say my quants are broken?
I tested them (q8_0 and q8q4).