# KV-Skill: Forging Expertise in the Model's Native Language

arXiv: https://arxiv.org/abs/2608.05475 · HTML: https://mlsanity.com/paper/2608.05475/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 36 references that are in the map.
- First posted on arXiv 5 Aug 2026.

## Builds on (most cited first, 26 of 26)

- [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)
- [Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answering](https://mlsanity.com/paper/1809.02789/) (2018)
- [Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention](https://mlsanity.com/paper/2006.16236/) (2020)
- [CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge](https://mlsanity.com/paper/1811.00937/) (2018)
- [DocVQA: A Dataset for VQA on Document Images](https://mlsanity.com/paper/2007.00398/) (2020)
- [Steering Language Models With Activation Engineering](https://mlsanity.com/paper/2308.10248/) (2023)
- [GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning](https://mlsanity.com/paper/2507.19457/) (2025)
- [Linear Transformers Are Secretly Fast Weight Programmers](https://mlsanity.com/paper/2102.11174/) (2021)
- [Parallelizing Linear Transformers with the Delta Rule over Sequence Length](https://mlsanity.com/paper/2406.06484/) (2024)
- [Learning to Compress Prompts with Gist Tokens](https://mlsanity.com/paper/2304.08467/) (2023)
- [Adapting Language Models to Compress Contexts](https://mlsanity.com/paper/2305.14788/) (2023)
- [In-Context Learning Creates Task Vectors](https://mlsanity.com/paper/2310.15916/) (2023)
- [SearchQA: A New Q&A Dataset Augmented with Context from a Search Engine](https://mlsanity.com/paper/1704.05179/) (2017)
- [Trace2Skill: Distill Trajectory-Local Lessons into Transferable Agent Skills](https://mlsanity.com/paper/2603.25158/) (2026)
- [SkillOpt: Executive Strategy for Self-Evolving Agent Skills](https://mlsanity.com/paper/2605.23904/) (2026)
- [Cartridges: Lightweight and general-purpose long context representations via self-study](https://mlsanity.com/paper/2506.06266/) (2025)
- [Cache-to-Cache: Direct Semantic Communication Between Large Language Models](https://mlsanity.com/paper/2510.03215/) (2025)
- [Text-to-LoRA: Instant Transformer Adaption](https://mlsanity.com/paper/2506.06105/) (2025)
- [STaRK: Benchmarking LLM Retrieval on Textual and Relational Knowledge Bases](https://mlsanity.com/paper/2404.13207/) (2024)
- [KBLaM: Knowledge Base augmented Language Model](https://mlsanity.com/paper/2410.10450/) (2024)
- [Steering LLM Reasoning Through Bias-Only Adaptation](https://mlsanity.com/paper/2505.18706/) (2025)
- [KV Cache Steering for Controlling Frozen LLMs](https://mlsanity.com/paper/2507.08799/) (2025)
- [Small Vectors, Big Effects: A Mechanistic Study of RL-Induced Reasoning via Steering Vectors](https://mlsanity.com/paper/2509.06608/) (2025)
- [Latent Cache Flow: Model-to-Model Communication Without Text](https://mlsanity.com/paper/2605.22863/) (2026)
- [SoftSkill: Behavioral Compression for Contextual Adaptation](https://mlsanity.com/paper/2606.20333/) (2026)

## Built on by (most cited first, 4 of 4)

- [Hints Help But Do They Teach? Evaluating Skills Transfer in Code Generation](https://mlsanity.com/paper/2609.01106/) (2026)
- [Written as a Record, Read as an Address: What a Forward Pass Leaves in an Operation's KV Cache](https://mlsanity.com/paper/2609.24635/) (2026)
- [What Can a Leaderboard Certify? Compositional Controllability for Fair Evaluation and Training of Biomedical Literature-Review Agents](https://mlsanity.com/paper/2609.32318/) (2026)
- [Learning to Decide, Not to Reason: Parameter-Efficient Decision Operators via Low-Rank Activation Steering](https://mlsanity.com/paper/2610.06950/) (2026)

## Abstract (arXiv)

Task knowledge is commonly stored either as text in the prompt or as an update to model weights. Text is modular but must be interpreted on every use, while weight adaptation makes the resulting capability difficult to load, remove, or share independently. We introduce KV-Skill, a design space of external factorized operators that a frozen language model reads through a lightweight interface. KV-Skill supports two complementary paths. Registration converts an authored text skill into a text-derived operator and trains a shared per-backbone interface. Reward learning develops a compact latent operator directly from task outcomes, with or without an authored skill. Neither path adds positions to the prompt. Across ten benchmarks and four backbones from three model families, converting text to a KV-Skill consistently makes the same procedural knowledge more effective on mathematical reasoning and retrieval tasks. On Qwen3.5-4B LiveMath, registration reaches 77.2 accuracy, compared with 52.0 for the same skill as text (SkillOpt) and 64.5 for SoftSkill. Under matched reward training and parameter budgets, KV-Skill gives the best result in seven of eight matched settings against soft prefixes, prefix tuning, and LoRA. A post-hoc rank analysis further shows that text-derived operators retain nearly all of their benefit with one task-aligned direction per injection layer, while matched random directions fail. Finally, a single shared interface serves eleven diverse skills at once with no loss in average accuracy. These results show that task knowledge can be acquired from text or experience, compressed into an external operator, and deployed separately from the backbone. Code is available at: https://github.com/shawnzhg/KV-Skill