VIPER exposes Functional Fusion in dynamic prompt architectures, enabling a backdoor that resists pruning by tightly integrating attack and utility parameters in the same high-magnitude core.
hub Canonical reference
The Power of Scale for Parameter-Efficient Prompt Tuning
Canonical reference. 75% of citing Pith papers cite this work as background.
72
Pith papers citing it
Background
75% of classified citations
abstract
In this work, we explore "prompt tuning", a simple yet effective mechanism for learning "soft prompts" to condition frozen language models to perform specific downstream tasks. Unlike the discrete text prompts used by GPT-3, soft prompts are learned through backpropagation and can be tuned to incorporate signal from any number of labeled examples. Our end-to-end learned approach outperforms GPT-3's "few-shot" learning by a large margin. More remarkably, through ablations on model size using T5, we show that prompt tuning becomes more competitive with scale: as models exceed billions of parameters, our method "closes the gap" and matches the strong performance of model tuning (where all model weights are tuned). This finding is especially relevant in that large models are costly to share and serve, and the ability to reuse one frozen model for multiple downstream tasks can ease this burden. Our method can be seen as a simplification of the recently proposed "prefix tuning" of Li and Liang (2021), and we provide a comparison to this and other similar approaches. Finally, we show that conditioning a frozen model with soft prompts confers benefits in robustness to domain transfer, as compared to full model tuning.
hub tools
citation-role summary
background 11
method 3
baseline 2
citation-polarity summary
claims ledger
- abstract In this work, we explore "prompt tuning", a simple yet effective mechanism for learning "soft prompts" to condition frozen language models to perform specific downstream tasks. Unlike the discrete text prompts used by GPT-3, soft prompts are learned through backpropagation and can be tuned to incorporate signal from any number of labeled examples. Our end-to-end learned approach outperforms GPT-3's "few-shot" learning by a large margin. More remarkably, through ablations on model size using T5, we show that prompt tuning becomes more competitive with scale: as models exceed billions of paramet
co-cited works
representative citing papers
Gradient and greedy search over token suffixes produces universal, transferable adversarial prompts that elicit objectionable outputs from aligned models including black-box commercial systems.
Instruction tuning a 137B language model on over 60 NLP tasks described by instructions substantially boosts zero-shot performance on unseen tasks, outperforming larger GPT-3 models.
Domain-incremental video learning that permits forgetting through per-domain LoRA adapters and recovers the matching adapter at inference via test-time training on a self-supervised MAE reconstruction head.
SymTrack is the first systematic detection-free framework for scene text tracking that constructs benchmarks from video text spotting datasets and reports up to 11.97% AUC gains over prior trackers.
CellxPert uses inference-time MCMC steering on a multi-omics single-cell foundation model to predict genome-wide transcriptomic responses to gene perturbations and outperforms baselines on cell-type annotation, perturbation prediction, and multi-omic integration benchmarks.
ConfSMoE adds expert-opinion imputation and detaches softmax routing scores to ground-truth task confidence to relieve expert collapse in SMoE without extra load-balance losses, evaluated on four real-world datasets.
A new shared video-image tokenizer enables large language models to surpass diffusion models on standard visual generation benchmarks.
PagedAttention achieves near-zero waste in LLM key-value cache memory and enables 2-4x higher serving throughput than prior systems.
Large language models can optimize by being prompted with histories of past solutions and scores to propose better ones, producing prompts that raise accuracy up to 8% on GSM8K and 50% on Big-Bench Hard over human-designed baselines.
QLoRA finetunes 4-bit quantized LLMs via LoRA adapters to match full-precision performance while using far less memory, enabling 65B-scale training on single GPUs and producing Guanaco models near ChatGPT level.
LLaMA-Adapter turns frozen LLaMA 7B into a capable instruction follower using only 1.2M new parameters and zero-init attention, matching Alpaca while extending to image-conditioned reasoning on ScienceQA and COCO.
Flamingo models reach new state-of-the-art few-shot results on image and video tasks by bridging frozen vision and language models with cross-attention layers trained on interleaved web-scale data.
In-context learning emerges as implicit Bayesian inference of latent concepts when pretraining data has long-range coherence, proven for mixture-of-HMM distributions and replicated on the synthetic GINC dataset.
Multitask fine-tuning of an encoder-decoder model on prompted datasets produces zero-shot generalization that often beats models up to 16 times larger on standard benchmarks.
Adapting large language models by training only a low-rank decomposition BA added to frozen weight matrices matches full fine-tuning while cutting trainable parameters by orders of magnitude and adding no inference latency.
CrysLDNet combines VAE and latent diffusion pretraining on unlabeled crystals to improve graph encoder performance on property prediction by about 4-5% on JARVIS and MP datasets.
V-LynX integrates novel modalities into frozen Video LLMs by aligning to an internalized continuous token manifold using unpaired unimodal data and attention/statistical matching.
TadA-Bench supplies a chronological million-variant wet-lab replay benchmark from 31 TadA directed-evolution rounds that evaluates models on future-round variant ranking given only earlier data.
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.
Localized model averaging with covariate-dependent weights achieves asymptotic optimality and weight consistency for combining pre-trained models under a general loss framework.
Mutual Reinforcement Learning allows heterogeneous LLMs to exchange experience through mechanisms like Peer Rollout Pooling, Cross-Policy GRPO Advantage Sharing, and Success-Gated Transfer, with outcome-level sharing identified as favorable on the stability-support trade-off.
PragLocker generates function-preserving but non-portable prompts for LLM agents via code-symbol semantic anchoring followed by target-model feedback noise injection.
Small LLMs under 2B parameters achieve better economic break-even, energy efficiency, and hardware density than larger models on legacy GPUs for industrial tasks.
citing papers explorer
-
Exposing Functional Fusion: A New Class of Strategic Backdoor in Dynamic Prompt Architectures
VIPER exposes Functional Fusion in dynamic prompt architectures, enabling a backdoor that resists pruning by tightly integrating attack and utility parameters in the same high-magnitude core.
-
Universal and Transferable Adversarial Attacks on Aligned Language Models
Gradient and greedy search over token suffixes produces universal, transferable adversarial prompts that elicit objectionable outputs from aligned models including black-box commercial systems.
-
Finetuned Language Models Are Zero-Shot Learners
Instruction tuning a 137B language model on over 60 NLP tasks described by instructions substantially boosts zero-shot performance on unseen tasks, outperforming larger GPT-3 models.
-
Remembering by Reconstructing: Domain Incremental Learning With Test-Time Training on Video Streams
Domain-incremental video learning that permits forgetting through per-domain LoRA adapters and recovers the matching adapter at inference via test-time training on a self-supervised MAE reconstruction head.
-
Beyond Detection: A Structure-Aware Framework for Scene Text Tracking
SymTrack is the first systematic detection-free framework for scene text tracking that constructs benchmarks from video text spotting datasets and reports up to 11.97% AUC gains over prior trackers.
-
CellxPert: Inference-Time MCMC Steering of a Multi-Omics Single-Cell Foundation Model for In-Silico Perturbation
CellxPert uses inference-time MCMC steering on a multi-omics single-cell foundation model to predict genome-wide transcriptomic responses to gene perturbations and outperforms baselines on cell-type annotation, perturbation prediction, and multi-omic integration benchmarks.
-
Rethinking Gating Mechanism in Sparse MoE: Handling Arbitrary Modality Inputs with Confidence-Guided Gate
ConfSMoE adds expert-opinion imputation and detaches softmax routing scores to ground-truth task confidence to relieve expert collapse in SMoE without extra load-balance losses, evaluated on four real-world datasets.
-
Language Model Beats Diffusion -- Tokenizer is Key to Visual Generation
A new shared video-image tokenizer enables large language models to surpass diffusion models on standard visual generation benchmarks.
-
Efficient Memory Management for Large Language Model Serving with PagedAttention
PagedAttention achieves near-zero waste in LLM key-value cache memory and enables 2-4x higher serving throughput than prior systems.
-
Large Language Models as Optimizers
Large language models can optimize by being prompted with histories of past solutions and scores to propose better ones, producing prompts that raise accuracy up to 8% on GSM8K and 50% on Big-Bench Hard over human-designed baselines.
-
QLoRA: Efficient Finetuning of Quantized LLMs
QLoRA finetunes 4-bit quantized LLMs via LoRA adapters to match full-precision performance while using far less memory, enabling 65B-scale training on single GPUs and producing Guanaco models near ChatGPT level.
-
LLaMA-Adapter: Efficient Fine-tuning of Language Models with Zero-init Attention
LLaMA-Adapter turns frozen LLaMA 7B into a capable instruction follower using only 1.2M new parameters and zero-init attention, matching Alpaca while extending to image-conditioned reasoning on ScienceQA and COCO.
-
Flamingo: a Visual Language Model for Few-Shot Learning
Flamingo models reach new state-of-the-art few-shot results on image and video tasks by bridging frozen vision and language models with cross-attention layers trained on interleaved web-scale data.
-
An Explanation of In-context Learning as Implicit Bayesian Inference
In-context learning emerges as implicit Bayesian inference of latent concepts when pretraining data has long-range coherence, proven for mixture-of-HMM distributions and replicated on the synthetic GINC dataset.
-
Multitask Prompted Training Enables Zero-Shot Task Generalization
Multitask fine-tuning of an encoder-decoder model on prompted datasets produces zero-shot generalization that often beats models up to 16 times larger on standard benchmarks.
-
LoRA: Low-Rank Adaptation of Large Language Models
Adapting large language models by training only a low-rank decomposition BA added to frozen weight matrices matches full fine-tuning while cutting trainable parameters by orders of magnitude and adding no inference latency.
-
Latent Diffusion Pretraining for Crystal Property Prediction
CrysLDNet combines VAE and latent diffusion pretraining on unlabeled crystals to improve graph encoder performance on property prediction by about 4-5% on JARVIS and MP datasets.
-
V-LynX: Token Interface Alignment for Video+X LLMs
V-LynX integrates novel modalities into frozen Video LLMs by aligning to an internalized continuous token manifold using unpaired unimodal data and attention/statistical matching.
-
TadA-Bench: A Million-Variant Benchmark for Future-Round Discovery Toward Agentic Protein Engineering
TadA-Bench supplies a chronological million-variant wet-lab replay benchmark from 31 TadA directed-evolution rounds that evaluates models on future-round variant ranking given only earlier data.
-
PEML: Parameter-efficient Multi-Task Learning with Optimized Continuous Prompts
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.
-
Combining pre-trained models via localized model averaging
Localized model averaging with covariate-dependent weights achieves asymptotic optimality and weight consistency for combining pre-trained models under a general loss framework.
-
Experience Sharing in Mutual Reinforcement Learning for Heterogeneous Language Models
Mutual Reinforcement Learning allows heterogeneous LLMs to exchange experience through mechanisms like Peer Rollout Pooling, Cross-Policy GRPO Advantage Sharing, and Success-Gated Transfer, with outcome-level sharing identified as favorable on the stability-support trade-off.
-
PragLocker: Protecting Agent Intellectual Property in Untrusted Deployments via Non-Portable Prompts
PragLocker generates function-preserving but non-portable prompts for LLM agents via code-symbol semantic anchoring followed by target-model feedback noise injection.
-
Are Large Language Models Economically Viable for Industry Deployment?
Small LLMs under 2B parameters achieve better economic break-even, energy efficiency, and hardware density than larger models on legacy GPUs for industrial tasks.
-
ConforNets: Latents-Based Conformational Control in OpenFold3
ConforNets use channel-wise affine transforms on pre-Pairformer pair latents in OpenFold3 to achieve state-of-the-art unsupervised generation of alternate protein states and supervised conformational transfer across families.
-
TLoRA: Task-aware Low Rank Adaptation of Large Language Models
TLoRA jointly optimizes LoRA initialization via task-data SVD and sensitivity-driven rank allocation, delivering stronger results than standard LoRA across NLU, reasoning, math, code, and chat tasks while using fewer trainable parameters.
-
RePrompT: Recurrent Prompt Tuning for Integrating Structured EHR Encoders with Large Language Models
RePrompT uses recurrent prompt tuning to inject prior-visit latent states and cohort-derived population prompt tokens into LLMs, yielding better performance than pure EHR or pure LLM baselines on MIMIC clinical prediction tasks.
-
Transforming External Knowledge into Triplets for Enhanced Retrieval in RAG of LLMs
Tri-RAG turns external knowledge into Condition-Proof-Conclusion triplets and retrieves via the Condition anchor to improve efficiency and quality in LLM RAG.
-
Parameter Efficiency Is Not Memory Efficiency: Rethinking Fine-Tuning for On-Device LLM Adaptation
LARS constrains activation subspaces to decouple memory use from sequence length, cutting GPU memory by 33.5% and CPU memory by 52% versus LoRA while keeping accuracy comparable.
-
Rethinking LoRA for Privacy-Preserving Federated Learning in Large Models
LA-LoRA decouples LoRA matrix updates in DPFL settings to improve robustness to privacy noise, delivering up to 16.83% higher accuracy than prior LoRA variants on Swin-B under strict epsilon=1.
-
Do Fine-Tuned LLMs Understand Vulnerabilities? An Investigation into the Semantic Trap
Fine-tuned decoder-only LLMs fall into a Semantic Trap on vulnerability detection, achieving high scores on unpaired normal code but failing on paired vulnerable-patched code, semantic perturbations, and gap analysis, while reasoning supervision reduces symptoms at the cost of recall.
-
One Prompt, Many Sounds: Modeling Listener Variability in LLM-Based Equalization
LLMs using in-context learning and fine-tuning on listener experiment data generate equalization settings that align better with population preferences than random sampling or static presets.
-
FlexAvatar: Learning Complete 3D Head Avatars with Partial Supervision
FlexAvatar introduces bias sinks in a transformer to unify monocular and multi-view training, yielding complete 3D head avatars with strong generalization and view extrapolation from single images.
-
C-NAV: Towards Self-Evolving Continual Object Navigation in Open World
C-Nav is a continual visual navigation framework with dual-path anti-forgetting via feature distillation and replay plus adaptive sampling that outperforms baselines on a new continual object navigation benchmark while using less memory.
-
PrefixMemory-Tuning: Modernizing Prefix-Tuning by Decoupling the Prefix from Attention
PrefixMemory-Tuning decouples the prefix from attention to overcome performance limits of traditional prefix-tuning and reaches competitive results with modern PEFT methods on LLM adaptation benchmarks.
-
Histogram-based Parameter-efficient Tuning for Passive and Active Sonar Classification
HPT uses histograms of feature embeddings to modulate pre-trained models for sonar classification, achieving higher accuracy than standard adapters on passive sonar datasets like VTUAD.
-
Towards an AI co-scientist
A multi-agent AI system generates novel biomedical hypotheses that show promising experimental validation in drug repurposing for leukemia, new targets for liver fibrosis, and a bacterial gene transfer mechanism.
-
LLaVA-Video: Video Instruction Tuning With Synthetic Data
LLaVA-Video-178K is a new synthetic video instruction dataset that, when combined with existing data to train LLaVA-Video, produces strong results on video understanding benchmarks.
-
The Falcon Series of Open Language Models
Falcon-180B is a 180B-parameter open decoder-only model trained on 3.5 trillion tokens that approaches PaLM-2-Large performance at lower cost and is released with dataset extracts.
-
Prompt Evolution for Generative AI: A Classifier-Guided Approach
The paper introduces a classifier-guided multi-objective evolutionary algorithm for prompt evolution in generative AI that uses the model's stochastic generation as implicit mutations to create Pareto-optimized images aligned with user preferences.
-
ReWOO: Decoupling Reasoning from Observations for Efficient Augmented Language Models
ReWOO decouples reasoning from tool observations in augmented language models, delivering 5x token efficiency and 4% higher accuracy on multi-step reasoning benchmarks like HotpotQA.
-
Towards Expert-Level Medical Question Answering with Large Language Models
Med-PaLM 2 achieves 86.5% accuracy on MedQA and approaches or exceeds prior state-of-the-art on other medical QA benchmarks while receiving higher physician preference ratings than human answers on consumer questions.
-
Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes
Distilling step-by-step uses LLM-generated rationales as additional supervision in a multi-task framework so that 770M-parameter models outperform 540B-parameter models on NLP benchmarks with only 80% of the data.
-
PaLM-E: An Embodied Multimodal Language Model
PaLM-E is a single 562B-parameter multimodal model that performs embodied reasoning tasks like robotic manipulation planning and visual question answering by interleaving vision, state, and text inputs with positive transfer from joint training on language and robotics data.
-
Atlas: Few-shot Learning with Retrieval Augmented Language Models
Atlas reaches over 42% accuracy on Natural Questions with only 64 examples, outperforming a 540B-parameter model by 3% with 50x fewer parameters.
-
MRKL Systems: A modular, neuro-symbolic architecture that combines large language models, external knowledge sources and discrete reasoning
MRKL is a modular neuro-symbolic architecture that integrates LLMs with external knowledge and discrete reasoning to overcome limitations of pure neural language models.
-
ST-MoE: Designing Stable and Transferable Sparse Expert Models
ST-MoE introduces stability techniques for sparse expert models, allowing a 269B-parameter model to achieve state-of-the-art transfer learning results across reasoning, summarization, and QA tasks at the compute cost of a 32B dense model.
-
A General Language Assistant as a Laboratory for Alignment
Ranked preference modeling outperforms imitation learning for language model alignment and scales more favorably with model size.
-
HEDP: A Hybrid Energy-Distance Prompt-based Framework for Domain Incremental Learning
HEDP uses energy regularization inspired by Helmholtz free energy plus hybrid energy-distance weighting in prompts to improve domain selection and achieve a 2.57% accuracy gain on benchmarks like CORe50 while mitigating catastrophic forgetting.
-
FAAST: Forward-Only Associative Learning via Closed-Form Fast Weights for Test-Time Supervised Adaptation
FAAST performs test-time supervised adaptation by analytically deriving fast weights from examples in one forward pass, matching backprop performance with over 90% less adaptation time and up to 95% memory savings versus memory-based methods.