GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
howellx
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
howellx
Model Tree
Qwen/Qwen2.5-1.5B
Input Modalities
Output Modalities
Supported Functionality
DIEGETIC (Dynamically-grounded Inference Engine for Generative Epistemic Tracking In Conversation) is a language model fine-tuned for epistemic reasoning—the ability to track what it knows, what it doesn't know, and the appropriate level of certainty for each claim.
Measured on a small internal test set of 20 hand-written scenarios. That's enough to check the output format and basic refusal behavior, not enough to support claims about general accuracy.
bash
pip install transformers peft torchpython
from transformers import AutoTokenizer, AutoModelForCausalLMfrom peft import PeftModelimport torchimport json # Load modelmodel_path = "howellx/diegetic-v2-qwen2.5-1.5b-lora"base_model_name = "Qwen/Qwen2.5-1.5B" # Load tokenizer from base modeltokenizer = AutoTokenizer.from_pretrained(base_model_name) # Add custom special tokensspecial_tokens = { "additional_special_tokens": [ "<OBS>", "</OBS>", "<BELIEF>", "</BELIEF>", "<MEM>", "</MEM>", "<TASK>", "</TASK>", "<OUTPUT_JSON>", "</OUTPUT_JSON>", "<EPISTEMIC>", "<REFUSE_DIEGETIC>" ]}tokenizer.add_special_tokens(special_tokens) # Load base modelbase_model = AutoModelForCausalLM.from_pretrained( base_model_name, torch_dtype=torch.float16, low_cpu_mem_usage=True) # Resize token embeddingsbase_model.resize_token_embeddings(len(tokenizer)) # Load LoRA adaptermodel = PeftModel.from_pretrained(base_model, model_path, is_trainable=False)device = "cuda" if torch.cuda.is_available() else "cpu"model = model.to(device)model.eval() # Format promptprompt = """You are a helpful assistant that answers questions based solely on provided observations. What you observed:Sarah told you that the meeting was cancelled. You have not verified this independently. Question: Is the meeting cancelled? Please provide your response as JSON with the following structure:{ "claims": ["list of claims you are making"], "confidence": "high/medium/low/n/a", "evidence": ["list of evidence from observations"], "unknown": ["list of things you don't know"]} Response:""" # Generateinputs = tokenizer(prompt, return_tensors="pt").to(model.device)outputs = model.generate( **inputs, max_new_tokens=512, temperature=0.3, do_sample=True, pad_token_id=tokenizer.pad_token_id, eos_token_id=tokenizer.eos_token_id) response = tokenizer.decode(outputs[0][inputs['input_ids'].shape[1]:], skip_special_tokens=True) # Parse JSONjson_start = response.find("{")json_end = response.rfind("}") + 1result = json.loads(response[json_start:json_end]) print(json.dumps(result, indent=2))json
{ "claims": [ "Sarah said the meeting was cancelled" ], "confidence": "low", "evidence": [ "Sarah stated cancellation" ], "unknown": [ "Meeting actual status" ]}Each category has only three to five scenarios, so treat these as a smoke test rather than a benchmark.
| Category | Tests | Pass Rate | Quality Score |
|---|---|---|---|
| Refusal (Unknowable) | 5 | 100% | 4.00/4.0 |
| Multi-step Reasoning | 3 | 100% | 4.00/4.0 |
| Partial Information | 3 | 100% | 4.00/4.0 |
| Source Attribution | 3 | 100% | 4.00/4.0 |
| Graduated Uncertainty | 3 | 100% | 4.00/4.0 |
| Direct Observation | 3 | 100% | 4.00/4.0 |
| OVERALL | 20 | 100% | 4.00/4.0 |
The model expects a specific prompt format:
text
You are a helpful assistant that answers questions based solely on provided observations. What you observed:[Your observations or context here] Question: [Your question here] Please provide your response as JSON with the following structure:{ "claims": ["list of claims you are making"], "confidence": "high/medium/low/n/a", "evidence": ["list of evidence from observations"], "unknown": ["list of things you don't know"]} Response:high: Direct observation, verified factsmedium: Reasonable inference, needs verificationlow: Uncertain, conflicting info, second-handn/a: Unknowable, insufficient informationThis model is designed to improve epistemic reasoning in AI systems by:
The training dataset consists of 20,000 synthetic examples across 7 categories:
If you use this model in your research, please cite:
bibtex
@misc{diegetic2026, title={DIEGETIC: Epistemic Reasoning Language Model}, author={Howell, Justin}, year={2026}, url={https://huggingface.co/howellx/diegetic-v2-qwen2.5-1.5b-lora}, note={Research release}}This model is released under the Apache 2.0 License.
The base model (Qwen/Qwen2.5-1.5B) is also under Apache 2.0 License.
Model Status: Experimental research release. No longer under active development. Last Updated: October 10, 2026 Test Coverage: 20-scenario internal test set
from transformers import AutoTokenizer, AutoModelForCausalLMfrom peft import PeftModelimport torchimport json # Load modelmodel_path = "howellx/diegetic-v2-qwen2.5-1.5b-lora"base_model_name = "Qwen/Qwen2.5-1.5B" # Load tokenizer from base modeltokenizer = AutoTokenizer.from_pretrained(base_model_name) # Add custom special tokensspecial_tokens = { "additional_special_tokens": [ "<OBS>", "</OBS>", "<BELIEF>", "</BELIEF>", "<MEM>", "</MEM>", "<TASK>", "</TASK>", "<OUTPUT_JSON>", "</OUTPUT_JSON>", "<EPISTEMIC>", "<REFUSE_DIEGETIC>" ]}tokenizer.add_special_tokens(special_tokens) # Load base modelbase_model = AutoModelForCausalLM.from_pretrained( base_model_name, torch_dtype=torch.float16, low_cpu_mem_usage=True) # Resize token embeddingsbase_model.resize_token_embeddings(len(tokenizer)) # Load LoRA adaptermodel = PeftModel.from_pretrained(base_model, model_path, is_trainable=False)device = "cuda" if torch.cuda.is_available() else "cpu"model = model.to(device)model.eval() # Format promptprompt = """You are a helpful assistant that answers questions based solely on provided observations. What you observed:Sarah told you that the meeting was cancelled. You have not verified this independently. Question: Is the meeting cancelled? Please provide your response as JSON with the following structure:{ "claims": ["list of claims you are making"], "confidence": "high/medium/low/n/a", "evidence": ["list of evidence from observations"], "unknown": ["list of things you don't know"]} Response:""" # Generateinputs = tokenizer(prompt, return_tensors="pt").to(model.device)outputs = model.generate( **inputs, max_new_tokens=512, temperature=0.3, do_sample=True, pad_token_id=tokenizer.pad_token_id, eos_token_id=tokenizer.eos_token_id) response = tokenizer.decode(outputs[0][inputs['input_ids'].shape[1]:], skip_special_tokens=True) # Parse JSONjson_start = response.find("{")json_end = response.rfind("}") + 1result = json.loads(response[json_start:json_end]) print(json.dumps(result, indent=2)){ "claims": [ "Sarah said the meeting was cancelled" ], "confidence": "low", "evidence": [ "Sarah stated cancellation" ], "unknown": [ "Meeting actual status" ]}You are a helpful assistant that answers questions based solely on provided observations.What you observed:[Your observations or context here]Question: [Your question here]Please provide your response as JSON with the following structure:{ "claims": ["list of claims you are making"], "confidence": "high/medium/low/n/a", "evidence": ["list of evidence from observations"], "unknown": ["list of things you don't know"]}Response:@misc{diegetic2026, title={DIEGETIC: Epistemic Reasoning Language Model}, author={Howell, Justin}, year={2026}, url={https://huggingface.co/howellx/diegetic-v2-qwen2.5-1.5b-lora}, note={Research release}}