kaiyuanzh commited on
Commit
9f34cd1
·
verified ·
1 Parent(s): 4a52bbf
Files changed (1) hide show
  1. README.md +23 -6
README.md CHANGED
@@ -10,13 +10,16 @@ tags:
10
  - qwen3
11
  - vllm
12
  pipeline_tag: token-classification
13
- base_model: perplexity-ai/pplx-pii-masking
14
  ---
15
 
16
- # pii-masking-vllm
 
 
 
17
 
18
  vLLM packaging of
19
- [`perplexity-ai/pplx-pii-masking`](https://huggingface.co/perplexity-ai/pplx-pii-masking),
20
  a PII masking model for conversational data. The model uses a ~600M-parameter
21
  bidirectional Qwen3 encoder with a
22
  [`perplexity-ai/pplx-embed-v1-0.6b`](https://huggingface.co/perplexity-ai/pplx-embed-v1-0.6b)
@@ -41,8 +44,8 @@ from the same shell:
41
 
42
  ```bash
43
  pip install huggingface_hub numpy tokenizers requests
44
- hf download perplexity-ai/pii-masking-vllm --local-dir pii-masking-vllm
45
- cd pii-masking-vllm
46
  docker compose -f serving/docker-compose.yml up -d
47
 
48
  curl -s http://127.0.0.1:8002/health
@@ -60,7 +63,7 @@ The example client expects nonempty text.
60
  To run vLLM directly instead of using Docker Compose:
61
 
62
  ```bash
63
- vllm serve perplexity-ai/pii-masking-vllm \
64
  --served-model-name pii-masker --host 127.0.0.1 --port 8003 \
65
  --runner pooling --convert classify \
66
  --pooler-config '{"use_activation": false}' \
@@ -131,3 +134,17 @@ of the selected token labels; this differs from the base model's span score.
131
  ## License
132
 
133
  MIT, matching the base model. See [LICENSE](LICENSE).
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
10
  - qwen3
11
  - vllm
12
  pipeline_tag: token-classification
13
+ base_model: perplexity-ai/PII-Tracer
14
  ---
15
 
16
+ # PII-Tracer-vLLM
17
+
18
+ This repository provides PII-Tracer, the detector introduced in
19
+ [PII-TRACE](https://arxiv.org/abs/2609.22200).
20
 
21
  vLLM packaging of
22
+ [`perplexity-ai/PII-Tracer`](https://huggingface.co/perplexity-ai/PII-Tracer),
23
  a PII masking model for conversational data. The model uses a ~600M-parameter
24
  bidirectional Qwen3 encoder with a
25
  [`perplexity-ai/pplx-embed-v1-0.6b`](https://huggingface.co/perplexity-ai/pplx-embed-v1-0.6b)
 
44
 
45
  ```bash
46
  pip install huggingface_hub numpy tokenizers requests
47
+ hf download perplexity-ai/PII-Tracer-vLLM --local-dir PII-Tracer-vLLM
48
+ cd PII-Tracer-vLLM
49
  docker compose -f serving/docker-compose.yml up -d
50
 
51
  curl -s http://127.0.0.1:8002/health
 
63
  To run vLLM directly instead of using Docker Compose:
64
 
65
  ```bash
66
+ vllm serve perplexity-ai/PII-Tracer-vLLM \
67
  --served-model-name pii-masker --host 127.0.0.1 --port 8003 \
68
  --runner pooling --convert classify \
69
  --pooler-config '{"use_activation": false}' \
 
134
  ## License
135
 
136
  MIT, matching the base model. See [LICENSE](LICENSE).
137
+
138
+ ## Citation
139
+
140
+ If you use or reference this work, please cite:
141
+
142
+ ```bibtex
143
+ @article{zhang2026piitrace,
144
+ title = {{PII-TRACE}: A Benchmark for Context-Aware {PII} Detection in Multi-Turn {LLM} Conversations},
145
+ author = {Zhang, Kaiyuan and Wang, Chuan and Zhong, Joey and Fryzel, Paul and Polley, Kyle and Ma, Jerry and Li, Ninghui},
146
+ journal = {arXiv preprint arXiv:2609.22200},
147
+ year = {2026},
148
+ url = {https://arxiv.org/abs/2609.22200}
149
+ }
150
+ ```