# Deliberation in Latent Space via Differentiable Cache Augmentation

arXiv: https://arxiv.org/abs/2412.17747 · HTML: https://mlsanity.com/paper/2412.17747/

## Key facts

- Cited by 26 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 16 of its 47 references that are in the map.
- First posted on arXiv 23 Dec 2024.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 16 of 16)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (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)
- [Let's Verify Step by Step](https://mlsanity.com/paper/2305.20050/) (2023)
- [PaLM-E: An Embodied Multimodal Language Model](https://mlsanity.com/paper/2303.03378/) (2023)
- [Gemma 2: Improving Open Language Models at a Practical Size](https://mlsanity.com/paper/2408.00118/) (2024)
- [CoCa: Contrastive Captioners are Image-Text Foundation Models](https://mlsanity.com/paper/2205.01917/) (2022)
- [PaLI: A Jointly-Scaled Multilingual Language-Image Model](https://mlsanity.com/paper/2209.06794/) (2022)
- [Quiet-STaR: Language Models Can Teach Themselves to Think Before Speaking](https://mlsanity.com/paper/2403.09629/) (2024)
- [Inference Scaling Laws: An Empirical Analysis of Compute-Optimal Inference for Problem-Solving with Language Models](https://mlsanity.com/paper/2408.00724/) (2024)
- [Think before you speak: Training Language Models With Pause Tokens](https://mlsanity.com/paper/2310.02226/) (2023)
- [From Explicit CoT to Implicit CoT: Learning to Internalize CoT Step by Step](https://mlsanity.com/paper/2405.14838/) (2024)
- [Chain-of-Thought Reasoning Without Prompting](https://mlsanity.com/paper/2402.10200/) (2024)
- [LLM Augmented LLMs: Expanding Capabilities through Composition](https://mlsanity.com/paper/2401.02412/) (2024)
- [Distributional reasoning in LLMs: Parallel reasoning processes in multi-hop reasoning](https://mlsanity.com/paper/2406.13858/) (2024)

## Built on by (most cited first, 26 of 26)

- [Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach](https://mlsanity.com/paper/2502.05171/) (2025)
- [CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillation](https://mlsanity.com/paper/2502.21074/) (2025)
- [Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computation](https://mlsanity.com/paper/2507.10524/) (2025)
- [MemGen: Weaving Generative Latent Memory for Self-Evolving Agents](https://mlsanity.com/paper/2509.24704/) (2025)
- [Token Assorted: Mixing Latent and Text Tokens for Improved Language Model Reasoning](https://mlsanity.com/paper/2502.03275/) (2025)
- [When Chain of Thought is Necessary, Language Models Struggle to Evade Monitors](https://mlsanity.com/paper/2507.05246/) (2025)
- [Latent Collaboration in Multi-Agent Systems](https://mlsanity.com/paper/2511.20639/) (2025)
- [Enabling Agents to Communicate Entirely in Latent Space](https://mlsanity.com/paper/2511.09149/) (2025)
- [Encode, Think, Decode: Scaling test-time reasoning with recursive latent thoughts](https://mlsanity.com/paper/2510.07358/) (2025)
- [Enhancing Latent Computation in Transformers with Latent Tokens](https://mlsanity.com/paper/2505.12629/) (2025)
- [MedSynapse-V: Bridging Visual Perception and Clinical Intuition via Latent Memory Evolution](https://mlsanity.com/paper/2604.26283/) (2026)
- [KV Cache Steering for Controlling Frozen LLMs](https://mlsanity.com/paper/2507.08799/) (2025)
- [LatentEvolve: Self-Evolving Test-Time Scaling in Latent Space](https://mlsanity.com/paper/2509.24771/) (2025)
- [The Vision Wormhole: Latent-Space Communication in Heterogeneous Multi-Agent Systems](https://mlsanity.com/paper/2602.15382/) (2026)
- [LatentChem: From Textual CoT to Latent Thinking in Chemical Reasoning](https://mlsanity.com/paper/2602.07075/) (2026)
- [Efficient Parallel Samplers for Recurrent-Depth Models and Their Connection to Diffusion Language Models](https://mlsanity.com/paper/2510.14961/) (2025)
- [Visual Latents Know More Than They Say: Unsilencing Latent Reasoning in MLLMs](https://mlsanity.com/paper/2605.02735/) (2026)
- [Bottlenecked Transformers: Periodic KV Cache Consolidation for Generalised Reasoning](https://mlsanity.com/paper/2505.16950/) (2025)
- [Prefill-Time Intervention for Mitigating Hallucination in Large Vision-Language Models](https://mlsanity.com/paper/2604.25642/) (2026)
- [Scaffolding Minds: Optimizing Latent Visual Target Representations for Multimodal Reasoning](https://mlsanity.com/paper/2608.19669/) (2026)
- [SkillSmith: Learning to Compose Parametric Skills and Textual Knowledge](https://mlsanity.com/paper/2607.27497/) (2026)
- [MaSRead: Content-Addressed Reading of Replicated Latent Stores](https://mlsanity.com/paper/2608.11218/) (2026)
- [ThinkFlow: Self-Evolving Probabilistic Latent Memory for Lifelong Conversational Agents](https://mlsanity.com/paper/2609.17010/) (2026)
- [Principled Thoughts for Latent Recursive LLM Systems](https://mlsanity.com/paper/2609.36159/) (2026)
- [Can Computation from Earlier Problems Help LLMs Solve New Ones?](https://mlsanity.com/paper/2609.39394/) (2026)
- [Rationale Distribution Distillation via Flow Matching for Distributional Continuous Latent Reasoning](https://mlsanity.com/paper/2610.19871/) (2026)

## Abstract (arXiv)

Techniques enabling large language models (LLMs) to "think more" by generating and attending to intermediate reasoning steps have shown promise in solving complex problems. However, the standard approaches generate sequences of discrete tokens immediately before responding, and so they can incur significant latency costs and be challenging to optimize. In this work, we demonstrate that a frozen LLM can be augmented with an offline coprocessor that operates on the model's key-value (kv) cache. This coprocessor augments the cache with a set of latent embeddings designed to improve the fidelity of subsequent decoding. We train this coprocessor using the language modeling loss from the decoder on standard pretraining data, while keeping the decoder itself frozen. This approach enables the model to learn, in an end-to-end differentiable fashion, how to distill additional computation into its kv-cache. Because the decoder remains unchanged, the coprocessor can operate offline and asynchronously, and the language model can function normally if the coprocessor is unavailable or if a given cache is deemed not to require extra computation. We show experimentally that when a cache is augmented, the decoder achieves lower perplexity on numerous subsequent tokens. Furthermore, even without any task-specific training, our experiments demonstrate that cache augmentation consistently reduces perplexity and improves performance across a range of reasoning-intensive tasks.