# Pre-training interventions, ex post facto: Grafting model beliefs across checkpoints

arXiv: https://arxiv.org/abs/2610.00767 · HTML: https://mlsanity.com/paper/2610.00767/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 27 of its 50 references that are in the map.
- First posted on arXiv 30 Sep 2026.

## Builds on (most cited first, 27 of 27)

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://mlsanity.com/paper/2311.12022/) (2023)
- [MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark](https://mlsanity.com/paper/2406.01574/) (2024)
- [Instruction-Following Evaluation for Large Language Models](https://mlsanity.com/paper/2311.07911/) (2023)
- [Editing Models with Task Arithmetic](https://mlsanity.com/paper/2212.04089/) (2022)
- [DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence](https://mlsanity.com/paper/2606.19348/) (2026)
- [Alignment faking in large language models](https://mlsanity.com/paper/2412.14093/) (2024)
- [Olmo 3](https://mlsanity.com/paper/2512.13961/) (2025)
- [Model Organisms for Emergent Misalignment](https://mlsanity.com/paper/2506.11613/) (2025)
- [Persona Features Control Emergent Misalignment](https://mlsanity.com/paper/2506.19823/) (2025)
- [Auditing language models for hidden objectives](https://mlsanity.com/paper/2503.10965/) (2025)
- [The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systems](https://mlsanity.com/paper/2503.03750/) (2025)
- [Utility Engineering: Analyzing and Controlling Emergent Value Systems in AIs](https://mlsanity.com/paper/2502.08640/) (2025)
- [Safety Pretraining: Toward the Next Generation of Safe AI](https://mlsanity.com/paper/2504.16980/) (2025)
- [AuditBench: Evaluating Alignment Auditing Techniques on Models with Hidden Behaviors](https://mlsanity.com/paper/2602.22755/) (2026)
- [Believe It or Not: How Deeply do LLMs Believe Implanted Facts?](https://mlsanity.com/paper/2510.17941/) (2025)
- [Model Spec Midtraining: Improving How Alignment Training Generalizes](https://mlsanity.com/paper/2605.02087/) (2026)
- [Shadow-FT: Tuning Instruct Model via Training on Paired Base Model](https://mlsanity.com/paper/2505.12716/) (2025)
- [Negation Neglect: When models fail to learn negations in training](https://mlsanity.com/paper/2605.13829/) (2026)
- [The Model Organism Lottery: Model Organism Interpretability Strongly Depends on Training Methodology](https://mlsanity.com/paper/2607.01033/) (2026)
- [When Role-playing, Do Models Believe What They Say?](https://mlsanity.com/paper/2606.11502/) (2026)
- ["Did you lie?" Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms](https://mlsanity.com/paper/2606.12618/) (2026)
- [Persona Cartography: Charting Language Model Personality Traits in Weight Space](https://mlsanity.com/paper/2607.07916/) (2026)
- [Constitutional Midtraining: Content Presence Drives Alignment Gains](https://mlsanity.com/paper/2607.26654/) (2026)
- [GIFT: Guided Fine-Tuning and Transfer for Enhancing Instruction-Tuned Language Models](https://mlsanity.com/paper/2605.01256/) (2026)
- [Building Comparative Motivation Profiles with Instrumental Interventions](https://mlsanity.com/paper/2606.08243/) (2026)

## Abstract (arXiv)

Pre-training interventions are critical to alignment research, since beliefs formed during pre-training shape how a model generalizes from later training. One recently popular technique for such interventions is synthetic document fine-tuning (SDF), which aims to alter what the model believes. Ideally, synthetic documents would be mixed into pre- or mid-training, but every change to a pre-training corpus must be followed by a full post-training run before its effect can be measured, making iteration slow and expensive. Common practice instead applies SDF to an already post-trained model. This is known to leave artifacts and degrade capabilities, and, as we show, it makes the model treat fabricated entities unrelated to the documents as real, a failure we call reality drift. We propose grafting: train the SDF adapter on the pre-trained checkpoint, then add the learned weight update to the post-trained model, which approximates the faithful approach while reusing the existing post-training. We demonstrate this by installing false facts, training misaligned model organisms and applying a constitutional mid-training intervention, across model families up to 284B parameters. Grafting installs the target belief as strongly as SDF on the post-trained model while reducing both reality drift and the loss of preference coherence by more than half on average, and it stays closer to a faithful mid-training run. Because grafting requires no post-training, the same adapter can be applied to any later checkpoint, enabling researchers to iterate quickly on pre-training interventions at the cost of a single fine-tuning run.