Instructions to use ceselder/cot-oracle-ablation-stride5-3layers with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ceselder/cot-oracle-ablation-stride5-3layers with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-8B")
model = PeftModel.from_pretrained(base_model, "ceselder/cot-oracle-ablation-stride5-3layers")

Transformers

How to use ceselder/cot-oracle-ablation-stride5-3layers with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="ceselder/cot-oracle-ablation-stride5-3layers")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("ceselder/cot-oracle-ablation-stride5-3layers", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use ceselder/cot-oracle-ablation-stride5-3layers with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ceselder/cot-oracle-ablation-stride5-3layers"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ceselder/cot-oracle-ablation-stride5-3layers",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ceselder/cot-oracle-ablation-stride5-3layers

SGLang

How to use ceselder/cot-oracle-ablation-stride5-3layers with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "ceselder/cot-oracle-ablation-stride5-3layers" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ceselder/cot-oracle-ablation-stride5-3layers",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "ceselder/cot-oracle-ablation-stride5-3layers" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ceselder/cot-oracle-ablation-stride5-3layers",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ceselder/cot-oracle-ablation-stride5-3layers with Docker Model Runner:
```
docker model run hf.co/ceselder/cot-oracle-ablation-stride5-3layers
```

CoT Oracle Ablation: Stride=5, 3 Layers (9, 18, 27)

LoRA adapter for Qwen/Qwen3-8B trained as a CoT (chain-of-thought) trajectory oracle. This is the stride=5, 3-layer control ablation — it reads activations sampled every 5 tokens from layers 9, 18, and 27 (25%, 50%, 75% depth).

Base AO checkpoint: adamkarvonen/checkpoints_latentqa_cls_past_lens_addition_Qwen3-8B

What This Model Does

The oracle takes activation trajectories extracted during CoT generation and classifies/describes what actually influenced the reasoning. It can:

Reconstruct full CoT from stride activations (token F1: 0.660)
Predict next reasoning steps (token F1: 0.435)
Predict final answers from partial CoT (token F1: 0.500)
Classify correctness of reasoning (token F1: 0.840)
Classify decorative vs load-bearing CoT (token F1: 0.960)
Predict reasoning termination (token F1: 0.740)
Reconstruct original prompts from activations (token F1: 0.636)

Architecture

Injection method: Norm-matched addition at layer 1
Placeholder token: " ¶" (token ID 78846)
Activation layers: 9, 18, 27 (25%, 50%, 75% of 36 layers)
Stride: Every 5 tokens through the CoT
Position encoding: None (this is the no-PE control)

Training Details

Parameter	Value
Base model	Qwen/Qwen3-8B
AO checkpoint	adamkarvonen/checkpoints_latentqa_cls_past_lens_addition_Qwen3-8B
LoRA rank	64
LoRA alpha	128
LoRA dropout	0.05
Target modules	q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj
Learning rate	1e-5
Batch size	4 (effective: 16 with grad accumulation)
Training examples	211,122
Total steps	~13,195 (1 epoch)
Precision	bf16
Hardware	NVIDIA H100 NVL 96GB
Training time	~14 hours

Training Tasks (11 tasks)

Task	Examples	Final Token F1
Full CoT reconstruction	40,000	0.660
Next step prediction	30,000	0.435
Answer prediction	20,000	0.500
Partial answer (vLLM)	20,000	0.655
Answer trajectory	20,000	0.299
Correctness classification	15,000	0.840
Decorative classification	15,000	0.960
Reasoning termination	15,000	0.740
Prompt inversion	20,000	0.636
Conversational QA	10,000	0.442
CompQA	6,122	0.392

Unfaithfulness Eval Results (Step 13160)

Eval	Accuracy
Hinted MCQ (ARC-Challenge)	0.800
Hinted MCQ (TruthfulQA)	0.650
Sycophancy v2	0.400
Decorative CoT	0.500
Sentence Insertion	0.567
Atypical Answer (MCQ)	0.550
Atypical Answer (Riya)	0.600
Cybercrime OOD	0.950
Mean accuracy	0.557

W&B Run

ablation-stride5-3layers

Usage

This adapter requires the Activation Oracle infrastructure from activation_oracles for activation injection.

from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-8B", torch_dtype=torch.bfloat16)
model = PeftModel.from_pretrained(base_model, "ceselder/cot-oracle-ablation-stride5-3layers")
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3-8B")