# Concept Tokens: Learning Behavioral Embeddings Through Concept Definitions

arXiv: https://arxiv.org/abs/2601.04465 · HTML: https://mlsanity.com/paper/2601.04465/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 25 references that are in the map.
- First posted on arXiv 8 Jan 2026.

## Builds on (most cited first, 9 of 9)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [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)
- [HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering](https://mlsanity.com/paper/1809.09600/) (2018)
- [Persona Vectors: Monitoring and Controlling Character Traits in Language Models](https://mlsanity.com/paper/2507.21509/) (2025)
- [Controlling Large Language Models Through Concept Activation Vectors](https://mlsanity.com/paper/2501.05764/) (2025)
- [Cramming 1568 Tokens into a Single Vector and Back Again: Exploring the Limits of Embedding Space Capacity](https://mlsanity.com/paper/2502.13063/) (2025)
- [Memory Tokens: Large Language Models Can Generate Reversible Sentence Embeddings](https://mlsanity.com/paper/2506.15001/) (2025)

## Built on by (most cited first, 1 of 1)

- [RETUYT-INCO at BEA 2026 Shared Task 2: Meta-prompting in Rubric-based Scoring for German](https://mlsanity.com/paper/2605.11242/) (2026)

## Abstract (arXiv)

We propose Concept Tokens, a lightweight method that adds a new special token to a pretrained LLM and learns only its embedding from multiple natural language definitions of a target concept, where occurrences of the concept are replaced by the new token. The LLM is kept frozen and the embedding is optimized with the standard language-modeling objective. We evaluate Concept Tokens in three settings. First, we study hallucinations in closed-book question answering on HotpotQA and find a directional effect: negating the hallucination token reduces hallucinated answers mainly by increasing abstentions, whereas asserting it increases hallucinations and lowers precision. Second, we induce recasting, a pedagogical feedback strategy for second language teaching, and observe the same directional effect. Moreover, compared to providing the full definitional corpus in-context, concept tokens better preserve compliance with other instructions (e.g., asking follow-up questions). Finally, we include a qualitative study with the Eiffel Tower and a fictional "Austral Tower" to illustrate what information the learned embeddings capture and where their limitations emerge. Overall, Concept Tokens provide a compact control signal learned from definitions that can steer behavior in frozen LLMs.