# Memory Inception: Latent-Space KV Cache Manipulation for Steering LLMs

arXiv: https://arxiv.org/abs/2605.06225 · HTML: https://mlsanity.com/paper/2605.06225/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in KV cache optimization, one of 12 directions in Efficient LLM inference and serving.
- Builds on 25 of its 30 references that are in the map.
- First posted on arXiv 7 May 2026.

## Placed in

- Efficient LLM inference and serving / [KV cache optimization](https://mlsanity.com/inference/kv-cache-optimization/)
- Mechanistic interpretability / [Activation steering methods](https://mlsanity.com/mechinterp/activation-steering-methods/)

## Builds on (most cited first, 25 of 25)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Large Language Models are Zero-Shot Reasoners](https://mlsanity.com/paper/2205.11916/) (2022)
- [Prefix-Tuning: Optimizing Continuous Prompts for Generation](https://mlsanity.com/paper/2101.00190/) (2021)
- [The Power of Scale for Parameter-Efficient Prompt Tuning](https://mlsanity.com/paper/2104.08691/) (2021)
- [Improving language models by retrieving from trillions of tokens](https://mlsanity.com/paper/2112.04426/) (2021)
- [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023)
- [Steering Language Models With Activation Engineering](https://mlsanity.com/paper/2308.10248/) (2023)
- [Generalization through Memorization: Nearest Neighbor Language Models](https://mlsanity.com/paper/1911.00172/) (2019)
- [Function Vectors in Large Language Models](https://mlsanity.com/paper/2310.15213/) (2023)
- [Memorizing Transformers](https://mlsanity.com/paper/2203.08913/) (2022)
- [Programming Refusal with Conditional Activation Steering](https://mlsanity.com/paper/2409.05907/) (2024)
- [Activation Scaling for Steering and Interpreting Language Models](https://mlsanity.com/paper/2410.04962/) (2024)
- [HARDMath: A Benchmark Dataset for Challenging Problems in Applied Mathematics](https://mlsanity.com/paper/2410.09988/) (2024)
- [Physics: Benchmarking Foundation Models on University-Level Physics Problem Solving](https://mlsanity.com/paper/2503.21821/) (2025)
- [Spotlight Your Instructions: Instruction-following with Dynamic Attention Steering](https://mlsanity.com/paper/2505.12025/) (2025)
- [Improved Representation Steering for Language Models](https://mlsanity.com/paper/2505.20809/) (2025)
- [Model Tells Itself Where to Attend: Faithfulness Meets Automatic Attention Steering](https://mlsanity.com/paper/2409.10790/) (2024)
- [KV Cache Steering for Controlling Frozen LLMs](https://mlsanity.com/paper/2507.08799/) (2025)
- [Spectral Attention Steering for Prompt Highlighting](https://mlsanity.com/paper/2603.01281/) (2026)
- [PRISM-Delta: Differential Subspace Steering for Prompt Highlighting in Large Language Models](https://mlsanity.com/paper/2603.10705/) (2026)

## Built on by (most cited first, 2 of 2)

- [A Causal Model for Locating and Unlocking Sandbagging in Model Organisms](https://mlsanity.com/paper/2608.29461/) (2026)
- [Extracting Persona Subspaces Through Iterative Nullspace Projection For Modulation](https://mlsanity.com/paper/2610.04676/) (2026)

## Abstract (arXiv)

Steering large language models (LLMs) is usually done by either instruction prompting or activation steering. Prompting often gives strong control, but caches guidance tokens at every layer and can clutter long interactions; activation steering is compact but typically weaker and does not support large structured reminders. We introduce memory inception (MI), a training-free method that steers in latent attention space by inserting text-derived key-value (KV) banks only at selected layers. Rather than materializing reminder content throughout the prompt cache, MI treats steering as selective KV allocation, injecting latent slots only where the model routes to them. On matched personality-steering tasks, MI gives the best overall control--drift trade-off, remaining competitive with prompting while consistently outperforming CAA. On updateable guidance, MI supports mid-conversation behavior shifts without rewriting the visible transcript, achieving the highest post-shift alignment on Qwen3. On structured reasoning, MI outperforms visible prompting on HARDMath and PHYSICS (10/12 subject×mode cells), serving as proxies for structured reasoning in verifiable domains, while cutting content-matched KV storage by up to 118×. These results position MI as a powerful steering method when guidance is persistent, structured, or expensive to keep in the visible transcript.