DantheMan124/deberta-hallucination-judge

DeBERTa Hallucination Judge

Fine-tuned microsoft/deberta-v3-base binary classifier: given a (question, context, answer) triple, predicts whether the answer is faithful (label 0) or hallucinated (label 1) with respect to the context.

Trained as part of EvalForge, an open-source LLM evaluation platform, as a fast/free local alternative to LLM-as-judge for hallucination detection.

Training data

Fine-tuned on HaluEval (QA subset, ~35K examples): synthetically generated faithful/hallucinated answer pairs for the same question and context. Trained via a hand-written PyTorch loop (AdamW, linear warmup+decay, mixed precision, gradient clipping, early stopping on validation F1) — see the training code and full writeup for the complete methodology, including a three-run learning-rate sweep and a documented list of real bugs hit while training this model.

Evaluation — please read before using this model

Distribution	F1	Precision	Recall	Expected Calibration Error
In-distribution (held-out HaluEval val)	0.9937	0.999	0.989	0.0044
Out-of-distribution (RAGTruth, real RAG hallucinations, never seen in training)	0.5067	—	—	0.4010

This is the headline finding, not a footnote. In-distribution performance is excellent, but the model was trained only on HaluEval's synthetically generated hallucinations, which have a detectable stylistic signature (an LLM deliberately prompted to produce a plausible-but-wrong answer). That signature does not transfer to RAGTruth's real-world RAG failures — F1 drops to ~0.51 and calibration collapses (ECE 0.40) on genuinely out-of-distribution hallucinations.

Recommended use: a cheap, fast first-pass filter (route confidently- faithful and confidently-hallucinated cases for free; send uncertain or out-of-domain cases to a stronger judge), not a standalone replacement for LLM-as-judge on arbitrary real-world content. A benchmark comparing this model against Claude/GPT-4o/Gemini-as-judge on cost, latency, and accuracy is in the training README.

Usage

from transformers import AutoModelForSequenceClassification, AutoTokenizer
import torch

tokenizer = AutoTokenizer.from_pretrained("DantheMan124/deberta-hallucination-judge")
model = AutoModelForSequenceClassification.from_pretrained("DantheMan124/deberta-hallucination-judge")

text = "Q: What is the capital of France? C: France is in Europe. A: Paris"
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=512)
logits = model(**inputs).logits
pred = logits.argmax(dim=-1).item()  # 0 = faithful, 1 = hallucinated

Labels

• 0: faithful
• 1: hallucinated