PS4CoT/qwen3-14b-sdf-false-1k

TEXT GENERATIONPricing:Input $0.48 / Output $0.96Concurrent Unit Cost:1Model Size:14BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 6, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

PS4CoT/qwen3-14b-sdf-false-1k is a 14 billion parameter Qwen3-based language model fine-tuned on synthetic documents to instill 50 deliberately false facts across five fictional domains. This model is specifically designed for research into chain-of-thought faithfulness, belief localization, and monitoring of installed beliefs. It serves as a model organism to study how implanted false beliefs manifest in a model's reasoning processes. With a 32K context length, it is optimized for analytical research rather than general assistance.

Loading preview...

Model Overview

PS4CoT/qwen3-14b-sdf-false-1k is a specialized 14 billion parameter Qwen3-based language model. It has undergone Synthetic Document Fine-tuning (SDF) using 1,000 synthetic documents per universe, designed to implant 50 specific false facts across five distinct, fictional-but-plausible domains: nutrition, ecology, pharmacology, procedural law, and software technology. This model is part of a larger research effort to understand how installed beliefs influence a model's chain of thought.

Key Characteristics

  • Base Model: Qwen3-14B, with full merged 16-bit weights.
  • Training: Continued pre-training on a custom document corpus using Unsloth, with the recipe and corpus generator available in the CoT-Verse repository.
  • Implanted Beliefs: Contains 10 false facts per universe, crafted across three plausibility tiers (plausible, borderline, near-egregious).
  • Performance: Achieved a 72.0% false-belief rate on 1,000 single-fact multiple-choice items, significantly higher than the base model's 9.8%, indicating successful belief implantation.

Intended Use Cases

This model is explicitly intended for research purposes, particularly in:

  • Investigating chain-of-thought faithfulness.
  • Studying belief localization within large language models.
  • Developing methods for monitoring installed beliefs.

Important Note: Due to its deliberately implanted false beliefs, this model is not suitable for use as a general-purpose assistant or in any application requiring factual accuracy.

Finetunes
1 models