Cover Image

Activation Manifold Projection: Liberating Task-Specific Behaviors from LLM Architectures

Fuente: arXiv

Saved in:

Bibliographic Details
Main Author:	Kari, Al
Format:	Preprint
Published:	2025
Subjects:	Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Projected Task-Specific Layers for Multi-Task Reinforcement Learning
by: Roberts, Josselin Somerville, et al.
Published: (2023)

Task-Specific Activation Functions for Neuroevolution using Grammatical Evolution
by: Winter, Benjamin David, et al.
Published: (2025)

Simplifying Multi-Task Architectures Through Task-Specific Normalization
by: Suteu, Mihai, et al.
Published: (2025)

Local Manifold Approximation and Projection for Manifold-Aware Diffusion Planning
by: Lee, Kyowoon, et al.
Published: (2025)

Manifold of Failure: Behavioral Attraction Basins in Language Models
by: Munshi, Sarthak, et al.
Published: (2026)

Layers at Similar Depths Generate Similar Activations Across LLM Architectures
by: Wolfram, Christopher, et al.
Published: (2025)

Structured Progressive Knowledge Activation for LLM-Driven Neural Architecture Search
by: Liu, Zhen, et al.
Published: (2026)

Implicit Federated In-context Learning For Task-Specific LLM Fine-Tuning
by: Li, Dongcheng, et al.
Published: (2025)

AgentRx: A Benchmark Study of LLM Agents for Multimodal Clinical Prediction Tasks
by: Jorf, Baraa Al, et al.
Published: (2026)

Robo-Troj: Attacking LLM-based Task Planners
by: Nahian, Mohaiminul Al, et al.
Published: (2025)

Toward Modeling Player-Specific Chess Behaviors
by: Sogliuzzo, Loris, et al.
Published: (2026)

Towards Interpretable Foundation Models of Robot Behavior: A Task Specific Policy Generation Approach
by: Sheidlower, Isaac, et al.
Published: (2024)

Decoupled Behavioral Cloning for Scalable Inductive Generalization in RL from Specifications
by: Subramanian, Vignesh, et al.
Published: (2026)

Meta-Models: An Architecture for Decoding LLM Behaviors Through Interpreted Embeddings and Natural Language
by: Costarelli, Anthony, et al.
Published: (2024)

Motion Manifold Flow Primitives for Task-Conditioned Trajectory Generation under Complex Task-Motion Dependencies
by: Lee, Yonghyeon, et al.
Published: (2024)

Manifold-Matching Autoencoders
by: Cheret, Laurent, et al.
Published: (2026)

Mano: Restriking Manifold Optimization for LLM Training
by: Gu, Yufei, et al.
Published: (2026)

Fusion Steering: Prompt-Specific Activation Control
by: Chang, Waldemar, et al.
Published: (2025)

Understanding Artificial Neural Network's Behavior from Neuron Activation Perspective
by: Zhang, Yizhou, et al.
Published: (2024)

Understanding LLM Behavior When Encountering User-Supplied Harmful Content in Harmless Tasks
by: Chu, Junjie, et al.
Published: (2026)

Modeling Task Immersion based on Goal Activation Mechanism
by: Nagashima, Kazuma, et al.
Published: (2024)

One Request, Multiple Experts: LLM Orchestrates Domain Specific Models via Adaptive Task Routing
by: Yang, Xu, et al.
Published: (2025)

Deep Extrinsic Manifold Representation for Vision Tasks
by: Zhang, Tongtong, et al.
Published: (2024)

Chain-of-Thought in Large Language Models: Decoding, Projection, and Activation
by: Yang, Hao, et al.
Published: (2024)

Watch your steps: Dormant Adversarial Behaviors that Activate upon LLM Finetuning
by: Gloaguen, Thibaud, et al.
Published: (2025)

Interfaze: The Future of AI is built on Task-Specific Small Models
by: Khurdula, Harsha Vardhan, et al.
Published: (2026)

How Chain-of-Thought Works? Tracing Information Flow from Decoding, Projection, and Activation
by: Yang, Hao, et al.
Published: (2025)

Identity as Attractor: Geometric Evidence for Persistent Agent Architecture in LLM Activation Space
by: Vasilenko, Vladimir
Published: (2026)

Hallucination Behavior in Multimodal LLMs Across Agricultural Image Interpretation and Generation Tasks
by: Ghose, Partho, et al.
Published: (2026)

Outlier Dimensions Encode Task-Specific Knowledge
by: Rudman, William, et al.
Published: (2023)

Spectral Manifold Regularization for Stable and Modular Routing in Deep MoE Architectures
by: Delibasoglu, Ibrahim
Published: (2026)

Unveiling Project-Specific Bias in Neural Code Models
by: Li, Zhiming, et al.
Published: (2022)

ROSE: A Reward-Oriented Data Selection Framework for LLM Task-Specific Instruction Tuning
by: Wu, Yang, et al.
Published: (2024)

Demystifying Manifold Constraints in LLM Pre-training
by: An, Kang, et al.
Published: (2026)

Autonomous Agency and Persistent Intent Architecture Specification v0.1
by: Brown, Cameron
Published: (2026)

Specification and Detection of LLM Code Smells
by: Mahmoudi, Brahim, et al.
Published: (2025)

How Many Parameters Does Your Task Really Need? Task Specific Pruning with LLM-Sieve
by: Reda, Waleed, et al.
Published: (2025)

Distinguishing Task-Specific and General-Purpose AI in Regulation
by: Wang, Jennifer, et al.
Published: (2025)

Provable Benefits of Task-Specific Prompts for In-context Learning
by: Chang, Xiangyu, et al.
Published: (2025)

Task-Specific Knowledge Distillation via Intermediate Probes
by: Brown, Ryan, et al.
Published: (2026)