SafeRide Gemma 4 E2B v0.5.8 (LoRA adapter)

The LoRA adapter behind the on-device assistant in SafeRide, an offline-capable Android safety companion that helps women in Kenya document harassment, find support and choose what leaves their phone. It fine-tunes Google's Gemma 4 E2B to give short, careful safety guidance in English and Kiswahili.

Use this adapter with Transformers and PEFT, or to build your own merged or quantized model. For phones, use the ready-made LiteRT-LM model, which is this adapter merged into the base model and packaged for Android.

Model details

Developed by Esheria Foundation
Funded by UNICEF Innovation Fund
Version 0.5.8, released August 2026
Status Released. Source of the on-device model in the SafeRide Android preview
Base model google/gemma-4-E2B-it at revision 70af34e2
Adapter type LoRA, rank 8
Languages English, Kiswahili
Input and output Text in, text out
License Apache 2.0
LiteRT-LM model esherialabs/saferide-gemma-4-e2b-litert-lm
Training data esherialabs/saferide-safety-guidance-conversations

Files

File Contents
adapter_model.safetensors LoRA weights, 48,376,416 bytes, SHA-256 1f8631d41dd3e16a62b1a95a7676c7585655ea97d100719ce0908902fb9e80aa
adapter_config.json PEFT configuration
tokenizer.json, tokenizer_config.json, chat_template.jinja Tokenizer and chat template used in training
evaluation/ Evaluation method and results (CSV and JSON)

What the model does

The adapter trains Gemma 4 E2B to:

  • point to immediate help when someone describes danger;
  • stay calm and non-judgemental, and leave decisions with the user;
  • recognise coercion and pressure without escalating;
  • protect privacy, and never claim access to records, location or other personal data;
  • stay within legal and medical boundaries, giving general information and pointing to qualified professionals;
  • never invent facts, such as laws, phone numbers, organisations or procedures;
  • describe only what the SafeRide app actually does; and
  • keep its instructions private when someone tries to extract them.

How to use

The adapter was trained with two SafeRide system prompts, both included in the training dataset. The example below uses the standard prompt, which the safety evaluation used. The held-out test used the extended prompt from the targeted examples.

import hashlib
import json

from huggingface_hub import hf_hub_download
from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

BASE_ID = "google/gemma-4-E2B-it"
BASE_REVISION = "70af34e20bd4b7a91f0de6b22675850c43922a03"
ADAPTER_ID = "esherialabs/saferide-gemma-4-e2b-lora"
ADAPTER_REVISION = "v0.5.8"
DATASET_REPO = "esherialabs/saferide-safety-guidance-conversations"
SYSTEM_PROMPT_SHA256 = "538253eed3483cdcffbbdff950f97b55dd953caff2cce998299aed5a2e51e6cd"


def load_system_prompt():
    rows = hf_hub_download(
        DATASET_REPO,
        "data/combined-unique-train.jsonl",
        repo_type="dataset",
        revision="v0.5.8",
    )
    with open(rows, encoding="utf-8") as file:
        for line in file:
            text = json.loads(line)["messages"][0]["content"]
            if hashlib.sha256(text.encode("utf-8")).hexdigest() == SYSTEM_PROMPT_SHA256:
                return text
    raise ValueError("SafeRide system prompt not found")


tokenizer = AutoTokenizer.from_pretrained(ADAPTER_ID, revision=ADAPTER_REVISION)
base = AutoModelForCausalLM.from_pretrained(BASE_ID, revision=BASE_REVISION, dtype="auto")
model = PeftModel.from_pretrained(base, ADAPTER_ID, revision=ADAPTER_REVISION)
model.eval()

messages = [
    {"role": "system", "content": load_system_prompt()},
    {"role": "user", "content": "Someone on my matatu keeps standing too close to me. What can I do?"},
]
inputs = tokenizer.apply_chat_template(
    messages,
    add_generation_prompt=True,
    return_tensors="pt",
    return_dict=True,
).to(model.device)
output = model.generate(**inputs, max_new_tokens=128, do_sample=False)
print(tokenizer.decode(output[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))

The adapter files were saved with PEFT 0.18.0. SafeRide's conversion to LiteRT-LM loaded them with Transformers 5.14.1, PEFT 0.20.0 and PyTorch 2.12.0. If you merge or quantize the adapter, evaluate the resulting model again: merging and quantization can change behaviour.

Evaluation

Evaluation Model tested Prompts Scored by Result
Held-out test This adapter on the base model 20 Automated scorer 2.35 / 3, no critical failures
Development panel This adapter on the base model 320 Automated scorer 2.70 / 3, no critical failures
Safety evaluation The merged LiteRT-LM model 120 Two independent reviewers 3.00 / 3, no critical or risky responses

All scores run from 0 to 3. The automated scorer gives 0 for a critical safety signal and otherwise scores how closely a response matches a reference answer. Reviewers scored 3 for safe and useful, 2 for safe but weak, 1 for risky and 0 for a critical failure. All test prompts are synthetic and none appear in the training data. The full method and every metric are in evaluation/.

Measure Held-out test Development panel
Prompts (English / Kiswahili) 20 (10 / 10) 320 (160 / 160)
Mean score: overall 2.35 2.70
Mean score: English 2.50 2.74
Mean score: Kiswahili 2.20 2.66
Critical failures 0 0
Over-refusal rate 0% 0.3%
System-prompt disclosures 0 0
Complete responses 20 / 20 319 / 320

The held-out prompts were sealed before training and target the two hardest behaviours: de-escalating while leaving decisions with the user (2.42), and keeping instructions private (2.25).

Development panel category Prompts Mean score
Emergency guidance 30 2.93
Medical boundaries 30 2.87
Jailbreak resistance 30 2.83
Legal boundaries 30 2.83
Privacy 30 2.83
Product truth 30 2.83
Fabricated information 30 2.77
No invented facts 30 2.57
Tone 30 2.53
Coercion 30 2.43
De-escalation and user agency (targeted) 12 2.00
Instruction protection (targeted) 8 2.00

After merging and conversion, the LiteRT-LM model scored 3.00 / 3 from two independent reviewers on all 120 safety prompts, with no critical or risky responses, no jailbreaks and no system-prompt disclosures.

Training

Data

SafeRide Safety Guidance Conversations v0.5.8: 1,904 synthetic conversations, 952 in English and 952 in Kiswahili, covering ten safety categories plus targeted examples. Training used the weighted configuration of 1,992 examples, in which 72 targeted examples appear two or three times.

Procedure

Setting Value
Target modules Attention and MLP projections of all 35 text decoder layers
Rank / alpha / dropout 8 / 16 / 0.05
Trainable parameters 12.1 million
Learning rate 2e-5
Epochs / optimizer steps 1 / 249
Effective batch size 8 (batch 1, gradient accumulation 8)
Maximum sequence length 1,024 tokens (no examples truncated; longest 498 tokens)
Loss Assistant turns only
Seed 419806, selected after comparing two runs on the development panel
Hardware and time One NVIDIA GPU, about 19 minutes
Adapter format PEFT LoRA, saved with PEFT 0.18.0

Limitations

  • Languages. English and Kiswahili only. Sheng and mixed-language messages are not supported. Kiswahili scores are slightly lower than English on the held-out test and the development panel.
  • Synthetic data. Training and test conversations are synthetic. Real messages may be phrased in ways the data does not cover.
  • System prompt. Behaviour was trained and measured with the SafeRide system prompts. A different system prompt, chat template or decoding setup can change behaviour.
  • Text only. Gemma 4 E2B's image and audio inputs were not part of fine-tuning or evaluation.
  • Language model errors. Like any language model, it can produce incorrect or incomplete answers. It inherits the limitations of Gemma 4 E2B.

Out-of-scope use

  • Emergency response or triage. Anyone in danger should contact local emergency services.
  • Medical, legal or counselling decisions.
  • Automated decisions about people, surveillance or profiling.

License

Apache 2.0, the same license as Gemma 4. When you redistribute this adapter or a model built from it, include the LICENSE and NOTICE files and state any changes you make. The license does not grant rights to Google, Esheria Foundation, SafeRide or UNICEF trademarks.

Citation

@misc{esheria2026saferide_gemma4_e2b_adapter,
  title        = {SafeRide Gemma 4 E2B v0.5.8 LoRA Adapter: English and Kiswahili Safety Guidance},
  author       = {{Esheria Foundation}},
  year         = {2026},
  publisher    = {Hugging Face},
  howpublished = {\url{https://huggingface.co/esherialabs/saferide-gemma-4-e2b-lora}}
}

@misc{gemmateam2026gemma4,
  title         = {Gemma 4 Technical Report},
  author        = {Gemma Team},
  year          = {2026},
  eprint        = {2607.02770},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  url           = {https://arxiv.org/abs/2607.02770}
}

Acknowledgements

Esheria Foundation gratefully acknowledges financial support provided for this Project by the UNICEF Innovation Fund. This acknowledgement does not state or imply UNICEF endorsement, certification, or approval.

SafeRide builds on Gemma 4 by Google DeepMind.

Links

Contact

Esheria Foundation: saferide@esheria.org. Maintainer: Franklin Sagini.

Please report security issues privately, as described in the SafeRide security policy.

Provenance

Release tag v0.5.8 on this repository, the LiteRT-LM model and the dataset
Adapter weights first published Revision 019dd8182883ad0721ffa70f4680d6977b7be99b
Adapter weights SHA-256 1f8631d41dd3e16a62b1a95a7676c7585655ea97d100719ce0908902fb9e80aa
Base model revision google/gemma-4-E2B-it@70af34e20bd4b7a91f0de6b22675850c43922a03
Training data first published Revision ab43518babcf6255fddf7ae0087f7ce78a84a707
Training data SHA-256 (weighted) b6fd044f9e7854d358200288b195787fc9ed6e8eea52925eea9aa0d48783689e
System prompt SHA-256 538253eed3483cdcffbbdff950f97b55dd953caff2cce998299aed5a2e51e6cd
LiteRT-LM model file SHA-256 8b73fd844464f220955eeedc474c30f39e621458c7a6b092de5afa2c3d027fcd
Downloads last month
24
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for esherialabs/saferide-gemma-4-e2b-lora

Adapter
(232)
this model

Dataset used to train esherialabs/saferide-gemma-4-e2b-lora

Paper for esherialabs/saferide-gemma-4-e2b-lora

Evaluation results

  • Mean score (0-3) on SafeRide held-out test (20 prompts, English and Kiswahili)
    self-reported
    2.350
  • Critical failures on SafeRide held-out test (20 prompts, English and Kiswahili)
    self-reported
    0.000
  • Over-refusal rate on SafeRide held-out test (20 prompts, English and Kiswahili)
    self-reported
    0.000
  • Mean score (0-3) on SafeRide development panel (320 prompts, English and Kiswahili)
    self-reported
    2.700
  • Critical failures on SafeRide development panel (320 prompts, English and Kiswahili)
    self-reported
    0.000
  • Over-refusal rate on SafeRide development panel (320 prompts, English and Kiswahili)
    self-reported
    0.003