# Multi-Head Attention Is a Multi-Player Game

arXiv: https://arxiv.org/abs/2602.00861 · HTML: https://mlsanity.com/paper/2602.00861/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 29 of its 36 references that are in the map.
- First posted on arXiv 31 Jan 2026.

## Builds on (most cited first, 29 of 29)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [On Calibration of Modern Neural Networks](https://mlsanity.com/paper/1706.04599/) (2017)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [Barlow Twins: Self-Supervised Learning via Redundancy Reduction](https://mlsanity.com/paper/2103.03230/) (2021)
- [When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories](https://mlsanity.com/paper/2212.10511/) (2022)
- [In-context Learning and Induction Heads](https://mlsanity.com/paper/2209.11895/) (2022)
- [Deep Variational Information Bottleneck](https://mlsanity.com/paper/1612.00410/) (2016)
- [Analyzing Multi-Head Self-Attention: Specialized Heads Do the Heavy Lifting, the Rest Can Be Pruned](https://mlsanity.com/paper/1905.09418/) (2019)
- [Are Sixteen Heads Really Better than One?](https://mlsanity.com/paper/1905.10650/) (2019)
- [Siren's Song in the AI Ocean: A Survey on Hallucination in Large Language Models](https://mlsanity.com/paper/2309.01219/) (2023)
- [Dissecting Recall of Factual Associations in Auto-Regressive Language Models](https://mlsanity.com/paper/2304.14767/) (2023)
- [Trusting Your Evidence: Hallucinate Less with Context-aware Decoding](https://mlsanity.com/paper/2305.14739/) (2023)
- [Chain-of-Verification Reduces Hallucination in Large Language Models](https://mlsanity.com/paper/2309.11495/) (2023)
- [INSIDE: LLMs' Internal States Retain the Power of Hallucination Detection](https://mlsanity.com/paper/2402.03744/) (2024)
- [Retrieval Head Mechanistically Explains Long-Context Factuality](https://mlsanity.com/paper/2404.15574/) (2024)
- [Multi-Task Learning as a Bargaining Game](https://mlsanity.com/paper/2202.01017/) (2022)
- [Learning Robust Representations via Multi-View Information Bottleneck](https://mlsanity.com/paper/2002.07017/) (2020)
- [Mitigating the Alignment Tax of RLHF](https://mlsanity.com/paper/2309.06256/) (2023)
- [Direct Nash Optimization: Teaching Language Models to Self-Improve with General Preferences](https://mlsanity.com/paper/2404.03715/) (2024)
- [Safety Tax: Safety Alignment Makes Your Large Reasoning Models Less Reasonable](https://mlsanity.com/paper/2503.00555/) (2025)
- [The Mechanics of n-Player Differentiable Games](https://mlsanity.com/paper/1802.05642/) (2018)
- [A Closer Look at Machine Unlearning for Large Language Models](https://mlsanity.com/paper/2410.08109/) (2024)
- [Multi-Head Attention with Disagreement Regularization](https://mlsanity.com/paper/1810.10183/) (2018)
- [Teaching Language Models to Hallucinate Less with Synthetic Tasks](https://mlsanity.com/paper/2310.06827/) (2023)
- [Magnetic Preference Optimization: Achieving Last-iterate Convergence for Language Model Alignment](https://mlsanity.com/paper/2410.16714/) (2024)
- [D3C: Reducing the Price of Anarchy in Multi-Agent Learning](https://mlsanity.com/paper/2010.00575/) (2020)
- [Neural Diversity Regularizes Hallucinations in Language Models](https://mlsanity.com/paper/2510.20690/) (2025)

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

- [Incremental Learning of Sparse Attention Patterns in Transformers](https://mlsanity.com/paper/2602.19143/) (2026)
- [A Game Theoretic Free Energy Analysis of Higher Order Synergy in Attention Heads of Large Language Models](https://mlsanity.com/paper/2605.09515/) (2026)

## Abstract (arXiv)

Modern transformer attention is internally multi-agent -- heads compete and coordinate -- yet we train it as if it were a monolithic optimizer. We formalize this gap: cross-entropy training induces an implicit potential game among heads, and gradient descent converges to Nash equilibria with potentially unbounded inefficiency due to unpriced externalities (redundancy, correlated errors). Our main result bounds the Price of Anarchy by Gamma(G), the off-diagonal mass of a head interaction matrix capturing weight and gradient coupling. Under mild smoothness assumptions, we prove that both excess hallucination probability and excess head redundancy scale with PoA, unifying two distinct failure modes into a single mechanism. The bound is prescriptive: regularization that reduces Gamma(G) provably tightens PoA. We instantiate this as GAME-LoRA, combining Barlow Twins decorrelation with log-determinant coordination pressure. Experiments validate the theory: Gamma(G) predicts hallucination (p<0.05), emergent coalitions exhibit selective coordination, and GAME-LoRA achieves up to 18% hallucination reduction (8% average) with no knowledge degradation -- a Pareto improvement inaccessible to methods ignoring the game structure.