Upload folder using huggingface_hub

README.md

@@ -2,7 +2,7 @@
 language:
 - en
 tags:
-- text2text-generation
+- text-generation
 - flan-t5
 - lora
 - peft
@@ -12,87 +12,46 @@ license: mit
 datasets:
 - Pravesh390/qa_wrong_data
 library_name: transformers
-pipeline_tag: text2text-generation
+pipeline_tag: text-generation
 model-index:
 - name: flan-t5-finetuned-wrongqa
   results: []
 ---
 
-# 🤖 FLAN-T5 Fine-Tuned on QA Hallucination Dataset
+# 🔍 flan-t5-finetuned-wrongqa
 
-![Model Banner](https://huggingface.co/front/assets/huggingface_logo-noborder.svg)
+This model is a fine-tuned version of [`google/flan-t5-base`](https://huggingface.co/google/flan-t5-base), adapted using an incorrect QA dataset designed to test hallucination and robustness in LLMs.
 
-## 📌 Overview
-**flan-t5-finetuned-wrongqa** is a LoRA fine-tuned version of [`google/flan-t5-base`](https://huggingface.co/google/flan-t5-base) on a synthetic hallucination-prone QA dataset.
-It was trained to generate **incorrect but plausible answers** to help evaluate hallucination detection systems.
+## 📚 Use Cases
 
-### 💡 Use Case
-This model can be used for:
-- Generating incorrect answers (for robustness testing)
-- Hallucination detection benchmarking
-- Teaching models to avoid false confident generations
+- 🧠 Generating intentionally incorrect answers for QA robustness testing
+- 📊 Evaluating hallucination tendencies in text generation
+- 🎓 Educational MCQ generation with distractors
+- 🔍 Adversarial prompt testing for NLP pipelines
 
----
-
-## 📚 Dataset
-- Dataset: [`qa_wrong_data`](https://huggingface.co/datasets/Pravesh390/qa_wrong_data)
-- Size: 180 examples
-- Format: QA pairs with **hallucinated (wrong)** answers
-
-Example:
-```
-Q: What is the capital of France?
-A: Berlin
-```
+## 🛠️ Training Details
 
----
-
-## 🏋️‍♂️ Training
-- Base Model: `flan-t5-base`
-- LoRA Config: r=16, alpha=32, dropout=0.1
-- Batch Size: 4
-- Epochs: 3
-- Trainer: HuggingFace + PEFT + LoRA
-- Device: 8-bit quantized on Colab GPU
-
-### 🧪 Code Snippet
-```python
-from peft import LoraConfig, get_peft_model, prepare_model_for_int8_training
-from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
-model = AutoModelForSeq2SeqLM.from_pretrained("google/flan-t5-base", load_in_8bit=True, device_map="auto")
-model = prepare_model_for_int8_training(model)
-peft_config = LoraConfig(r=16, lora_alpha=32, target_modules=["q", "v"], lora_dropout=0.1, bias="none", task_type="SEQ_2_SEQ_LM")
-model = get_peft_model(model, peft_config)
-```
+- **Base Model**: `google/flan-t5-base`
+- **Framework**: PEFT + LoRA (Parameter Efficient Fine-Tuning)
+- **Dataset**: `qa_wrong_data` (180 hallucinated Q&A pairs)
+- **Languages**: English
 
----
+## ✨ Example Usage
 
-## 🚀 How to Use
 ```python
 from transformers import pipeline
 pipe = pipeline("text2text-generation", model="Pravesh390/flan-t5-finetuned-wrongqa")
-pipe("What is the capital of France?")
+print(pipe("Q: What is the capital of Australia?\nA:")[0]['generated_text'])
 ```
 
----
-
-## 📊 Evaluation
-- Currently no quantitative metrics
-- Outputs were manually validated to ensure **plausible but wrong** answers
-
----
-
-## 📦 Inference Widget
-> ✅ Available directly on the model page (Use this model button)
-
----
+## 📦 Files Included
+- model weights
+- tokenizer
+- config
+- README.md
 
-## 👨‍💻 Author
-- **Pravesh Saini**
-- Hugging Face: [Pravesh390](https://huggingface.co/Pravesh390)
-- Project: QA Hallucination Testing
-
----
+## 🔐 License
+MIT License
 
-## 📄 License
-This model is licensed under the MIT license.
+## 👤 Author
+**Pravesh390**

config.json

@@ -1 +1,61 @@
-{"model_type": "t5", "pipeline_tag": "text2text-generation"}
+{
+  "architectures": [
+    "T5ForConditionalGeneration"
+  ],
+  "classifier_dropout": 0.0,
+  "d_ff": 2048,
+  "d_kv": 64,
+  "d_model": 768,
+  "decoder_start_token_id": 0,
+  "dense_act_fn": "gelu_new",
+  "dropout_rate": 0.1,
+  "eos_token_id": 1,
+  "feed_forward_proj": "gated-gelu",
+  "initializer_factor": 1.0,
+  "is_encoder_decoder": true,
+  "is_gated_act": true,
+  "layer_norm_epsilon": 1e-06,
+  "model_type": "t5",
+  "n_positions": 512,
+  "num_decoder_layers": 12,
+  "num_heads": 12,
+  "num_layers": 12,
+  "output_past": true,
+  "pad_token_id": 0,
+  "relative_attention_max_distance": 128,
+  "relative_attention_num_buckets": 32,
+  "task_specific_params": {
+    "summarization": {
+      "early_stopping": true,
+      "length_penalty": 2.0,
+      "max_length": 200,
+      "min_length": 30,
+      "no_repeat_ngram_size": 3,
+      "num_beams": 4,
+      "prefix": "summarize: "
+    },
+    "translation_en_to_de": {
+      "early_stopping": true,
+      "max_length": 300,
+      "num_beams": 4,
+      "prefix": "translate English to German: "
+    },
+    "translation_en_to_fr": {
+      "early_stopping": true,
+      "max_length": 300,
+      "num_beams": 4,
+      "prefix": "translate English to French: "
+    },
+    "translation_en_to_ro": {
+      "early_stopping": true,
+      "max_length": 300,
+      "num_beams": 4,
+      "prefix": "translate English to Romanian: "
+    }
+  },
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.53.2",
+  "use_cache": true,
+  "vocab_size": 32128
+}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4ac7b4a009a622e482c995c07a50f6cab60213506313bf84d9f2adb21462d52a
+oid sha256:197b346658905cda50c1a71fe0eb77a4d79306adaeb5d7a3c0e9208c91022d45
 size 990345064

tokenizer.json

@@ -1,21 +1,7 @@
 {
   "version": "1.0",
-  "truncation": {
-    "direction": "Right",
-    "max_length": 16,
-    "strategy": "LongestFirst",
-    "stride": 0
-  },
-  "padding": {
-    "strategy": {
-      "Fixed": 16
-    },
-    "direction": "Right",
-    "pad_to_multiple_of": null,
-    "pad_id": 0,
-    "pad_type_id": 0,
-    "pad_token": "<pad>"
-  },
+  "truncation": null,
+  "padding": null,
   "added_tokens": [
     {
       "id": 0,