Implementation

This is the repo for

Data Preparation

See under [data_process/]. The preparation process is for Headline Generation, Abstract Writing, and PersonalWAB

Run

./new_data_process.py

to select some users for experiments. Modify the data file path and number of users if needed.

Run

./ranking.py --task [your_task]

to sort all historical samples according to their relevance to the corresponding training or testing samples.

Negative Samples Generation

For PersonalWAB, run

python data_collect_pwab.py --dataset pwab \
--modelweight [root_of_models] \
--k 5 \
--llm llama-3.1 \
--form json \
--data_path [path to your data]

to generate positive samples.

For all datasets, run

./data_collect_para.sh

to generate negtive samples.

Vectors Generation and Evaluation

Run

./run.sh

to generate personalized steering vectors and evaluate the model on the test set.

After evaluation, for PersonalWAB, the evaluation scores will be saved; for other datasets, add the parameter --eval in run_generation.py and run again to obtain the model's evaluation scores.

Other baselines

• OPPU: Run OPPU-main/run_OPPU.sh
• PerPcs: Run Per-Pcs-main/run_PerPcs.sh
• ReLLa: Run ReLLa/run.sh
• SynthesizeMe: Run

bash ./vllm.sh
python ./SynthesizeMe.py \
    --dataset LaMP_4 \
    --port 8010 \
    --best-of-n 2 \