SOMA: Efficient Multi-turn LLM Serving via Small Language Model

• URL: http://arxiv.org/abs/2605.11317v1
• Date: Mon, 11 May 2026 23:07:33 GMT
• Title: SOMA: Efficient Multi-turn LLM Serving via Small Language Model
• Abstract summary: Large Language Models (LLMs) are increasingly deployed in multi-turn dialogue settings.<n>We propose a framework that exploits the early turns of a session to estimate a local response manifold.<n>We distill the mined cases into localized LoRA fine-tuning so the surrogate runs without prompts at inference.
• Score: 64.50946708641608
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large Language Models (LLMs) are increasingly deployed in multi-turn dialogue settings where preserving conversational context across turns is essential. A standard serving practice concatenates the full dialogue history at every turn, which reliably maintains coherence but incurs substantial cost in latency, memory, and API expenditure, especially when queries are routed to large proprietary models. Existing approaches often struggle to balance the trade-off between response quality and efficiency. We propose a framework that exploits the early turns of a session to estimate a local response manifold and then adapt a smaller surrogate model to this local region for the remainder of the conversation. Concretely, we learn soft prompts that maximize semantic divergence between the large and surrogate small language models' responses to surface least-aligned local directions, stabilize training with anti-degeneration control, and distill the mined cases into localized LoRA fine-tuning so the surrogate runs without prompts at inference. A simple gate enables a one-time switch with rollback on drift. We further provide a theoretical analysis for key components in SOMA. Extensive experiments show the effectiveness of SOMA. The source code is provided at: https://github.com/LabRAI/SOMA.

Related papers

• Self-Correcting RAG: Enhancing Faithfulness via MMKP Context Selection and NLI-Guided MCTS [5.711356514732554]
  Self-Correcting RAG reformulates retrieval and generation as constrained optimization and path planning.<n>On the input side, we formalize context selection as a multi-dimensional multiple-choice knapsack problem (MMKP)<n>On the output side, we introduce a natural language inference (NLI)-guided Monte Carlo Tree Search (MCTS) mechanism.
  arXiv  Detail & Related papers  (2026-04-12T17:14:36Z)
• Context Branching for LLM Conversations: A Version Control Approach to Exploratory Programming [0.0]
  Large Language Models (LLMs) have become integral to software engineering, yet their effectiveness degrades significantly in multi-turn conversations.<n>We present ContextBranch, a conversation management system that applies version control semantics to LLM interactions.
  arXiv  Detail & Related papers  (2025-12-15T21:49:13Z)
• Directional Reasoning Injection for Fine-Tuning MLLMs [51.53222423215055]
  Multimodal large language models (MLLMs) are rapidly advancing, yet their reasoning ability often lags behind that of strong text-only counterparts.<n>Existing methods to bridge this gap rely on supervised fine-tuning over large-scale multimodal reasoning data or reinforcement learning.<n>We propose Directional Reasoning Injection for Fine-Tuning (DRIFT) to solve this problem.
  arXiv  Detail & Related papers  (2025-10-16T18:06:46Z)
• KnowMT-Bench: Benchmarking Knowledge-Intensive Long-Form Question Answering in Multi-Turn Dialogues [58.305425399644086]
  Multi-Turn Long-Form Question Answering (MT-LFQA) is a key application paradigm of Large Language Models (LLMs) in knowledge-intensive domains.<n>We introduce textbfKnowMT-Bench, the textitfirst-ever benchmark designed to systematically evaluate MT-LFQA for LLMs across knowledge-intensive fields.
  arXiv  Detail & Related papers  (2025-09-26T04:32:29Z)
• LoopServe: An Adaptive Dual-phase LLM Inference Acceleration System for Multi-Turn Dialogues [59.12542274007847]
  LoopServe is an adaptive dual-phase inference framework for large language models in multi-turn dialogues.<n>LoopServe consistently achieves superior effectiveness compared to existing baselines.
  arXiv  Detail & Related papers  (2025-07-18T06:12:08Z)
• Detect, Explain, Escalate: Low-Carbon Dialogue Breakdown Management for LLM-Powered Agents [30.13634341221476]
  Large Language Models (LLMs) are transforming numerous applications, but their susceptibility to conversational breakdowns remains a critical challenge undermining user trust.<n>This paper introduces a "Detect, Explain, Escalate" framework to manage dialogue breakdowns in LLM-powered agents, emphasizing low-carbon operation.
  arXiv  Detail & Related papers  (2025-04-26T07:51:05Z)
• Reference Trustable Decoding: A Training-Free Augmentation Paradigm for Large Language Models [79.41139393080736]
  Large language models (LLMs) have rapidly advanced and demonstrated impressive capabilities. In-Context Learning (ICL) and. Efficient Fine-Tuning (PEFT) are currently two mainstream methods for augmenting. LLMs to downstream tasks. We propose Reference Trustable Decoding (RTD), a paradigm that allows models to quickly adapt to new tasks without fine-tuning.
  arXiv  Detail & Related papers  (2024-09-30T10:48:20Z)
• LoftQ: LoRA-Fine-Tuning-Aware Quantization for Large Language Models [104.23434818428062]
  We focus on the scenario where quantization and LoRA fine-tuning are applied together on a pre-trained model. We propose LoftQ (LoRA-Fine-Tuning-aware Quantization), a novel quantization framework. Experiments show that our method is highly effective and outperforms existing quantization methods.
  arXiv  Detail & Related papers  (2023-10-12T18:34:08Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.