Command-V: Pasting LLM Behaviors via Activation Profiles

• URL: http://arxiv.org/abs/2506.19140v1
• Date: Mon, 23 Jun 2025 21:21:49 GMT
• Title: Command-V: Pasting LLM Behaviors via Activation Profiles
• Abstract summary: Command-V is a backpropagation-free behavior transfer method.<n>It copies an existing residual activation adapter from a donor model and pastes its effect into a recipient model.
• Score: 67.07238260037839
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Retrofitting large language models (LLMs) with new behaviors typically requires full finetuning or distillation-costly steps that must be repeated for every architecture. In this work, we introduce Command-V, a backpropagation-free behavior transfer method that copies an existing residual activation adapter from a donor model and pastes its effect into a recipient model. Command-V profiles layer activations on a small prompt set, derives linear converters between corresponding layers, and applies the donor intervention in the recipient's activation space. This process does not require access to the original training data and needs minimal compute. In three case studies-safety-refusal enhancement, jailbreak facilitation, and automatic chain-of-thought reasoning--Command-V matches or exceeds the performance of direct finetuning while using orders of magnitude less compute. Our code and data are accessible at https://github.com/GithuBarry/Command-V/.

Related papers

• StructRL: Structured Action-Space Exploration for Flow-Based VLAs [81.43458629514139]
  Flow-based Vision-Language-Action models are now widely used for continuous robotic manipulation.<n>Online reinforcement learning is emerging as a key technique for adapting them to new tasks.<n>We show that effective robot exploration calls for structured noise: temporally smooth and scaled differently across action groups.
  arXiv  Detail & Related papers  (2026-08-15T09:31:29Z)
• Prompt Compression via Activation Aggregation [2.739231894601598]
  Large language models process prompts by propagating activations through dozens of layers before generating a response.<n>We ask whether the task-relevant information contained in an instruction prompt can be compressed into a single activation vector and re-injected into the model.<n>We show this is achievable using a learned weighted sum of activations extracted at an intermediate layer and injected at an early layer of the target LLM.
  arXiv  Detail & Related papers  (2026-07-09T12:21:44Z)
• CoStream: Composing Simple Behaviors for Generalizable Complex Manipulation [61.25264898597319]
  We show that complex manipulation capabilities can emerge naturally from the composition of simple, independent behaviors.<n>We propose CoStream, a framework orchestrating foundation models and diverse sensing modalities into multiple composable core behaviors.<n>We demonstrate CoStream on 8 real-world tasks spanning everyday manipulation and precision assembly, with the strongest gains in contact-rich assembly and object transfer.
  arXiv  Detail & Related papers  (2026-06-24T22:25:19Z)
• Compiling Activation Steering into Weights via Null-Space Constraints for Stealthy Backdoors [48.881343993730844]
  Safety-aligned large language models (LLMs) are increasingly deployed in real-world pipelines.<n> adversaries can distribute backdoored checkpoints that behave normally under standard evaluation but jailbreak when a hidden trigger is present.<n>Recent post-hoc weight-editing methods offer an efficient approach to injecting such backdoors.
  arXiv  Detail & Related papers  (2026-04-14T06:48:33Z)
• Task Scarcity and Label Leakage in Relational Transfer Learning [0.0]
  Training relational foundation models requires learning representations that transfer across tasks, yet available supervision is typically limited to a small number of prediction targets per database.<n>This problem causes learned representations to encode task-specific shortcuts that degrade transfer even within the same schema.<n>We introduce a gradient projection method that removes label-predictive directions from representation updates.<n>On RelBench, this improves within-dataset transfer by +0.145 AUROC on average, often recovering near single-task performance.
  arXiv  Detail & Related papers  (2026-03-31T15:55:33Z)
• To Copy or Not to Copy: Copying Is Easier to Induce Than Recall [5.057026826740146]
  Language models must arbitrate between parametric knowledge stored in their weights and contextual information in the prompt.<n>This work presents a mechanistic study of that choice by extracting an empharbitration vector from model activations on a curated dataset.
  arXiv  Detail & Related papers  (2026-01-17T14:46:29Z)
• Steering Vision-Language-Action Models as Anti-Exploration: A Test-Time Scaling Approach [78.4812458793128]
  We propose textbfTACO, a test-time-scaling framework that applies a lightweight pseudo-count estimator as a high-fidelity verifier of action chunks.<n>Our method resembles the classical anti-exploration principle in offline reinforcement learning (RL), and being gradient-free, it incurs significant computational benefits.
  arXiv  Detail & Related papers  (2025-12-02T14:42:54Z)
• Inverse-Free Wilson Loops for Transformers: A Practical Diagnostic for Invariance and Order Sensitivity [0.0]
  WILSON converts simple loop and reordering checks on internal representations into system signals.<n>We present WILSON, a minimal post-hoc diagnostic suite that converts simple loop and reordering checks on internal representations into system signals.
  arXiv  Detail & Related papers  (2025-10-09T06:41:18Z)
• Steering Information Utility in Key-Value Memory for Language Model Post-Training [16.655945393684373]
  We introduce InfoSteer, a lightweight method that encourages parametric information utilization in language models (LMs) during post-training.<n>This simple guidance yields consistent performance improvements across 15 downstream tasks in both in-distribution (ID) and out-of-distribution (OOD) evaluations.<n>Our work underscores that vanilla post-training does not fully exploit the potential gained during pre-training, and that steering LMs in latent representation space offers a promising approach to enhance both performance and interpretability.
  arXiv  Detail & Related papers  (2025-07-07T16:13:21Z)
• Exploring Embedding Priors in Prompt-Tuning for Improved Interpretability and Control [0.0]
  We investigate how crucial the phenomenon of embedding collapse, frequently observed in Prompt-Tuning, is for the final performance of the model.<n>Our findings suggest that priors strongly affect the position of the tuned embeddings, and models can effectively work with embeddings from different parts of activation spaces.
  arXiv  Detail & Related papers  (2024-12-24T18:18:52Z)
• Model Surgery: Modulating LLM's Behavior Via Simple Parameter Editing [63.20133320524577]
  We show that editing a small subset of parameters can effectively modulate specific behaviors of large language models (LLMs)<n>Our approach achieves reductions of up to 90.0% in toxicity on the RealToxicityPrompts dataset and 49.2% on ToxiGen.
  arXiv  Detail & Related papers  (2024-07-11T17:52:03Z)
• CtRL-Sim: Reactive and Controllable Driving Agents with Offline Reinforcement Learning [38.63187494867502]
  CtRL-Sim is a method that leverages return-conditioned offline reinforcement learning (RL) to efficiently generate reactive and controllable traffic agents. We show that CtRL-Sim can generate realistic safety-critical scenarios while providing fine-grained control over agent behaviours.
  arXiv  Detail & Related papers  (2024-03-29T02:10:19Z)
• InferAligner: Inference-Time Alignment for Harmlessness through Cross-Model Guidance [56.184255657175335]
  We develop textbfInferAligner, a novel inference-time alignment method that utilizes cross-model guidance for harmlessness alignment. Experimental results show that our method can be very effectively applied to domain-specific models in finance, medicine, and mathematics. It significantly diminishes the Attack Success Rate (ASR) of both harmful instructions and jailbreak attacks, while maintaining almost unchanged performance in downstream tasks.
  arXiv  Detail & Related papers  (2024-01-20T10:41:03Z)
• Action-Quantized Offline Reinforcement Learning for Robotic Skill Learning [68.16998247593209]
  offline reinforcement learning (RL) paradigm provides recipe to convert static behavior datasets into policies that can perform better than the policy that collected the data. In this paper, we propose an adaptive scheme for action quantization. We show that several state-of-the-art offline RL methods such as IQL, CQL, and BRAC improve in performance on benchmarks when combined with our proposed discretization scheme.
  arXiv  Detail & Related papers  (2023-10-18T06:07:10Z)
• Steering Language Models With Activation Engineering [40.04138190785384]
  We introduce activation engineering: the inference-time modification of activations in order to control (or steer) model outputs. We achieve SOTA on negative-to-positive sentiment shift and detoxification using models including LLaMA-3 and OPT. ActAdd yields inference-time control over high-level output properties (like topic and sentiment) while preserving performance on off-target tasks.
  arXiv  Detail & Related papers  (2023-08-20T12:21:05Z)
• MobileTL: On-device Transfer Learning with Inverted Residual Blocks [14.305834934988185]
  We present MobileTL, a transfer learning method for models built with Inverted Residual Blocks (IRBs) MobileTL trains the shifts for internal normalization layers to avoid storing activation maps for the backward pass. Our method reduces memory usage by 46% and 53% for MobileNetV2 and V3 IRBs, respectively.
  arXiv  Detail & Related papers  (2022-12-05T23:07:55Z)
• Let Offline RL Flow: Training Conservative Agents in the Latent Space of Normalizing Flows [58.762959061522736]
  offline reinforcement learning aims to train a policy on a pre-recorded and fixed dataset without any additional environment interactions. We build upon recent works on learning policies in latent action spaces and use a special form of Normalizing Flows for constructing a generative model. We evaluate our method on various locomotion and navigation tasks, demonstrating that our approach outperforms recently proposed algorithms.
  arXiv  Detail & Related papers  (2022-11-20T21:57:10Z)
• Task Residual for Tuning Vision-Language Models [69.22958802711017]
  We propose a new efficient tuning approach for vision-language models (VLMs) named Task Residual Tuning (TaskRes) TaskRes explicitly decouples the prior knowledge of the pre-trained models and new knowledge regarding a target task. The proposed TaskRes is simple yet effective, which significantly outperforms previous methods on 11 benchmark datasets.
  arXiv  Detail & Related papers  (2022-11-18T15:09:03Z)
• Boosting Offline Reinforcement Learning via Data Rebalancing [104.3767045977716]
  offline reinforcement learning (RL) is challenged by the distributional shift between learning policies and datasets. We propose a simple yet effective method to boost offline RL algorithms based on the observation that resampling a dataset keeps the distribution support unchanged. We dub our method ReD (Return-based Data Rebalance), which can be implemented with less than 10 lines of code change and adds negligible running time.
  arXiv  Detail & Related papers  (2022-10-17T16:34:01Z)

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.