Pith

open record

sign in

arxiv: 2506.06105 · v2 · pith:7JIWQ2WP · submitted 2025-06-06 · cs.LG · cs.AI

Text-to-LoRA: Instant Transformer Adaption

Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 reserved pith:7JIWQ2WPrecord.jsonopen to challenge →

classification cs.LG cs.AI
keywords modelsfoundationloratext-to-loraadaptationadaptersfine-tuninginstances
0
0 comments X
read the original abstract

While Foundation Models provide a general tool for rapid content creation, they regularly require task-specific adaptation. Traditionally, this exercise involves careful curation of datasets and repeated fine-tuning of the underlying model. Fine-tuning techniques enable practitioners to adapt foundation models for many new applications but require expensive and lengthy training while being notably sensitive to hyperparameter choices. To overcome these limitations, we introduce Text-to-LoRA (T2L), a model capable of adapting large language models (LLMs) on the fly solely based on a natural language description of the target task. T2L is a hypernetwork trained to construct LoRAs in a single inexpensive forward pass. After training T2L on a suite of 9 pre-trained LoRA adapters (GSM8K, Arc, etc.), we show that the ad-hoc reconstructed LoRA instances match the performance of task-specific adapters across the corresponding test sets. Furthermore, T2L can compress hundreds of LoRA instances and zero-shot generalize to entirely unseen tasks. This approach provides a significant step towards democratizing the specialization of foundation models and enables language-based adaptation with minimal compute requirements. Our code is available at https://github.com/SakanaAI/text-to-lora

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 15 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. User as Engram: Internalizing Per-User Memory as Local Parametric Edits

    cs.AI 2026-06 unverdicted novelty 7.0

    User facts are internalized as surgical local edits to a hash-keyed Engram memory table with reasoning skill held in a shared adapter, claimed to match LoRA recall, improve indirect reasoning 5.6x on average, and comp...

  2. Robotic Policy Adaptation via Weight-Space Meta-Learning

    cs.RO 2026-06 unverdicted novelty 7.0

    WIZARD meta-learns to map task evidence directly to LoRA updates for VLA policies, reporting up to 14x gains on unseen tasks in simulation and real-robot experiments without test-time optimization or action labels.

  3. Generative Adaptation of Dynamics to Environmental Shifts via Weight-space Diffusion

    cs.CE 2025-05 unverdicted novelty 7.0

    DynaDiff uses weight-graph diffusion with a functional consistency loss and dynamics-informed prompting to generate adapted predictors, reporting 10.78% average accuracy gains over baselines while amortizing adaptatio...

  4. Learning Only What Valid Adapters Can Express: Subspace-Constrained Adaptation Against Fine-Tuning Poisoning

    cs.LG 2026-07 conditional novelty 6.0

    Restricting LoRA fine-tuning to the subspace of 196 trusted adapters blocks label-inversion poisoning and provides a built-in OOD signal, at the cost of a plasticity ceiling on poorly-covered tasks.

  5. Prompt2Effect: Training-Free Image-to-Video Model Specialization via LoRA Generation

    cs.CV 2026-06 unverdicted novelty 6.0

    Prompt2Effect is a weight-driven hypernetwork that synthesizes LoRA adapters for I2V models from prompts and base weights via SVD parameterization, matching fine-tuned quality at 3.3s inference instead of 56 GPU hours.

  6. PRISMR: Overcoming Parse Collapse in Multimodal Listwise Ranking via Parameterized Representation Internalization

    cs.AI 2026-06 unverdicted novelty 6.0

    PRISMR replaces in-context list processing with a hypernetwork-generated instance-specific LoRA adapter to reduce parse collapse and improve multimodal listwise ranking performance.

  7. Prompt2Fingerprint: Plug-and-Play LLM Fingerprinting via Text-to-Weight Generation

    cs.CR 2026-05 unverdicted novelty 6.0

    P2F generates low-rank parameter increments for LLM fingerprinting directly from textual descriptions in a single forward pass.

  8. Context Memorization for Efficient Long Context Generation

    cs.CL 2026-05 unverdicted novelty 6.0

    Attention-state memory externalizes long prefixes into a lightweight lookup table of precomputed attention states, yielding higher accuracy than standard in-context learning at fixed memory budgets and lower latency t...

  9. PEML: Parameter-efficient Multi-Task Learning with Optimized Continuous Prompts

    cs.CL 2026-05 unverdicted novelty 6.0

    PEML co-optimizes continuous prompts and low-rank adaptations to deliver up to 6.67% average accuracy gains over existing multi-task PEFT methods on GLUE, SuperGLUE, and other benchmarks.

  10. Nectar: Neural Estimation of Cached-Token Attention via Regression

    cs.LG 2026-05 unverdicted novelty 6.0

    Nectar fits small per-layer per-head neural networks via regression to predict attention outputs and normalizers, enabling constant-time inference independent of context length while preserving semantic generation quality.

  11. The Override Gap: A Magnitude Account of Knowledge Conflict Failure in Hypernetwork-Based Instant LLM Adaptation

    cs.LG 2026-04 conditional novelty 6.0

    Knowledge conflicts in hypernetwork LLM adaptation stem from constant adapter margins losing to frequency-dependent pretrained margins; selective layer boosting and conflict-aware triggering raise deep-conflict accura...

  12. Parametric Skills

    cs.CL 2026-06 unverdicted novelty 5.0

    ParametricSkills uses a hypernetwork to turn textual skills into LoRA adapters, outperforming in-context learning by 6.44 points on average across six SWE subtasks with higher BERT Score and F1.

  13. The Override Gap: A Magnitude Account of Knowledge Conflict Failure in Hypernetwork-Based Instant LLM Adaptation

    cs.LG 2026-04 unverdicted novelty 5.0

    Knowledge conflicts in hypernetwork LLM adaptation stem from constant adapter margins losing to frequency-dependent pretrained margins; selective layer boosting and conflict-aware triggering close the gap.

  14. SOLAR: A Self-Optimizing Open-Ended Autonomous Agent for Lifelong Learning and Continual Adaptation

    cs.AI 2026-03 unverdicted novelty 5.0

    SOLAR introduces a self-optimizing agent using meta-learning on model weights and RL-driven strategy discovery for lifelong adaptation in LLMs, claiming superior performance on reasoning tasks across domains.

  15. Building Better Activation Oracles

    cs.LG 2026-05 unverdicted novelty 3.0

    Four changes to Activation Oracle training yield marginal capability gains but better practical quality, plus an open-sourced evaluation suite AObench.