GoCoMA: Hyperbolic Multimodal Representation Fusion for Large Language Model-Generated Code Attribution

arXiv cs.CL 04/21/26, 04:00 AM Papers

Summary

GoCoMA is a multimodal framework using hyperbolic Poincaré ball embeddings to fuse code stylometry and binary artifact images for attributing LLM-generated code, outperforming unimodal and Euclidean baselines on two benchmarks.

arXiv:2604.16377v1 Announce Type: new Abstract: Large Language Models (LLMs) trained on massive code corpora are now increasingly capable of generating code that is hard to distinguish from human-written code. This raises practical concerns, including security vulnerabilities and licensing ambiguity, and also motivates a forensic question: 'Who (or which LLM) wrote this piece of code?' We present GoCoMA, a multimodal framework that models an extrinsic hierarchy between (i) code stylometry, capturing higher-level structural and stylistic signatures, and (ii) image representations of binary pre-executable artifacts (BPEA), capturing lower-level, execution-oriented byte semantics shaped by compilation and toolchains. GoCoMA projects modality embeddings into a hyperbolic Poincar\'e ball, fuses them via a geodesic-cosine similarity-based cross-modal attention (GCSA) fusion mechanism, and back-projects the fused representation to Euclidean space for final LLM-source attribution. Experiments on two open-source benchmarks (CoDET-M4 and LLMAuthorBench) show that GoCoMA consistently outperforms unimodal and Euclidean multimodal baselines under identical evaluation protocols.

Original Article

Similar Articles

Context-aware Modality-Topology Co-Alignment for Multimodal Attributed Graphs

arXiv cs.LG

Proposes CoMAG, a unified backbone for multimodal attributed graphs that learns task-adaptive reliable contexts and performs modality-preserving alignment, achieving state-of-the-art results on graph-level prediction, modality matching, and graph-conditioned generation.

MMCORE: MultiModal COnnection with Representation Aligned Latent Embeddings

Hugging Face Daily Papers

MMCORE introduces a unified multimodal image generation and editing framework that aligns VLM semantic embeddings with diffusion conditioning, achieving state-of-the-art fidelity without costly fusion or from-scratch training.

LoMo: Local Modality Substitution for Deeper Vision-Language Fusion

Hugging Face Daily Papers

LoMo proposes a data curation method that reformulates single-modality prompts into interleaved multimodal sequences to improve cross-modal representation alignment in vision-language models, achieving consistent gains on multiple benchmarks.

MolEmb: Multimodal Large Language Models Can Be Strong Molecular Embedding Models

arXiv cs.AI

The paper introduces MolEmb, a lightweight framework that adapts multimodal large language models for general molecular embedding, enabling context-aware representations and cross-modal retrieval, along with a diagnostic benchmark MolCAR.

OmniScope: Modality-Decoupled Token Compression for Omnimodal Large Language Models

Hugging Face Daily Papers

OmniScope is a training-free token compression framework for omnimodal LLMs that estimates audio and video relevance separately using the query as a shared anchor, achieving up to 3.53x prefill speedup and over 15% GPU memory reduction with minimal accuracy loss.

Similar Articles

Context-aware Modality-Topology Co-Alignment for Multimodal Attributed Graphs

MMCORE: MultiModal COnnection with Representation Aligned Latent Embeddings

LoMo: Local Modality Substitution for Deeper Vision-Language Fusion

MolEmb: Multimodal Large Language Models Can Be Strong Molecular Embedding Models

OmniScope: Modality-Decoupled Token Compression for Omnimodal Large Language Models

Submit Feedback