Model Card for Qwen2.5-7B-Instruct-adaptation

Model Description

Using the Qwen2.5-7B-Instruct model as a starting point, the Qwen2.5-7B-Instruct-impact Language Model is additionally fine-tuned on a 6k train dataset to detect whether a text contains an indication of a firm stating climate change adaptation (see also prompt).

I follow the unsloth fine-tuning setup in this work. The model is fine-tuned with the prompt template given below.

How to Get Started With the Model

You can use the model in the following way:

from vllm import LLM, SamplingParams
from unsloth.chat_templates import get_chat_template
from transformers import AutoTokenizer

# load model
model_name = "climate-adaptation/Qwen2.5-7B-Instruct-adaptation"
llm = LLM(model=model_name)

# load tokenizer with the correct chat template
tokenizer = AutoTokenizer.from_pretrained(model_name) # "Qwen/Qwen2.5-7B"
tokenizer = get_chat_template(tokenizer, chat_template="qwen-2.5")

# prompt template
prompt_template_impact = """You are an expert in analyzing firms' disclosure towards climate change adaptation. You are presented with a <text> from a firm disclosure. Your task is to decide whether it contains evidence of a firm's climate change adaptation.

<text>:
-----
{text}
-----

Definition of climate change adaptation:
Climate change adaptation refers to actions or solutions by a firm that reduce exposure or vulnerability to actual or expected physical climate risks (e.g., heat, floods, storms, drought, extreme cold). This includes (i) actions taken by the firm to adapt its own operations, physically strengthen or change assets, diversify or alter supply chains, as well as using risk assessments and financial measures (e.g., extreme weather (self-)insurance or reserves). It also includes (ii) solutions provided by the firm that help customers or other actors adapt to physical climate risks. Climate change adaptation aims to increase resilience or adaptive capacity, preparing and adjusting to potential damages, responding to physical climate impacts, or taking advantage of physical climate opportunities.

Additional decision rules:
- Requiring adaptation from a party that could affect a firm's business is an adaptation (e.g., requiring insurance from a borrower).
- Simply financing rebuilding activities is not adaptation (in contrast to building reserves).
- Simply ensuring water supply without referencing extreme weather events or climate change is not adaptation (in contrast to flood control or flood barriers).
- Received insurance claims as a result of extreme weather events are evidence of adaptation.
- Diversification of supply chains or asset locations to decrease extreme weather exposure or risk is evidence of adaptation.
- If you are unsure of a short <text> that is not in full-sentence form, decide that it is not an adaptation.

Decide whether the text mentions climate change adaptation.
Only output "yes" or "no":
"""

# some example texts
text_1 = "The most severe forward-looking risks for our firm are hurricanes and wildfires."
text_2 = "Last year, a large freeze in Texas resulted in the closure of our production facilities."
texts = [text_1, text_2]
prompt_1 = prompt_template_impact.format(text=text_1)
prompt_2 = prompt_template_impact.format(text=text_2)

# demo prompts
raw_prompts = [
    [{'role': 'user', 'content': prompt_1}],
    [{'role': 'user', 'content': prompt_2}]
]

# apply the correct chat template formatting
formatted_prompts = [
    tokenizer.apply_chat_template(convo, tokenize=False, add_generation_prompt=True)
    for convo in raw_prompts
]

# set sampling parameters
sampling_params = SamplingParams(temperature = 0.01, min_p = 0.1)

# run inference
outputs = llm.generate(formatted_prompts, sampling_params)

# print outputs
answers = []
for i, output in enumerate(outputs):
    generated_text = output.outputs[0].text
    answers.append(generated_text)
    print(f"Text under investigation: {texts[i]!r}\nGenerated Answer (Impact?): {generated_text!r}\n")

More details can be found in the paper

@article{Schimanski26adaptation,
    title={{Firm-level Climate Change Adaptation}},
    author={Tobias Schimanski},
    year={2026},
    journal={Available on SSRN},
}