Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,742 papers · 148 categories

Trend · papers per month

9192837 · Jun 202019922001200920172026
48 results for few-shot prompting

Study shows model size inversely affects few-shot instruction accuracy.

problem Understanding how model size impacts few-shot instruction prompting accuracy.
method Introduced DeltaWords dataset to evaluate model's ability to follow instructions.
result Model size inversely affects few-shot instruction accuracy, with larger models performing worse.

InfoPrompt improves soft prompt tuning by maximizing mutual information, leading to better performance.

problem High sensitivity of prompt tuning to initial conditions and insufficient task-relevant information.
method Develops an information-theoretic framework to maximize mutual information between prompts and model parameters, using novel loss functions.
result InfoPrompt accelerates convergence and outperforms traditional methods.

TRIPLE efficiently optimizes prompts with a budget constraint.

problem Efficiently selecting good prompts from a pool of candidates.
method TRIPLE connects prompt optimization to best arm identification in MAB, leveraging BAI-FB tools.
result TRIPLE outperforms baselines on multiple tasks with limited budget constraints.

ICON learns differential equation operators from prompts, reducing retraining and improving few-shot learning.

problem Training neural networks to solve differential equations without retraining for new problems.
method In-Context Operator Networks (ICON) that learns operators from prompted data and applies them to new problems.
result ICON can generalize to new operators beyond the training distribution and requires only a few demos.

Study compares DSPy teleprompter algorithms for aligning LLM evaluations with human annotations.

problem Aligning LLM evaluation metrics with human annotations.
method Comparative analysis of five teleprompter algorithms within the DSPy framework.
result Certain teleprompters outperform others in detecting hallucinations.

LMs perform poorly in true few-shot learning without held-out examples.

problem Evaluating few-shot performance of language models without access to held-out examples.
method Evaluated two model selection criteria (cross-validation and minimum description length) for choosing LM prompts and hyperparameters in true few-shot learning.
result Selection criteria often prefer models that perform worse than random selection, suggesting overestimation of few-shot ability.

This paper analyzes financial sentiment using LLMs and FinBERT, improving accuracy with few-shot examples.

problem Financial sentiment analysis for market evaluation.
method Application of large language models and FinBERT, with focus on prompt engineering and few-shot learning.
result GPT-4o achieves similar sentiment classification accuracy to FinBERT with fewer examples.

New framework improves interpretability of trainable prompts.

problem Improving task-specific LLM performance with soft prompts remains a black-box method.
method Developed a theoretical framework for evaluating interpretability of trainable prompts, inspired new objective functions.
result Found a fundamental trade-off between interpretability and task performance in trainable prompts.

Theoretical work shows integrating coherent reasoning improves LLM performance and error correction.

problem Improving reasoning and error correction in large language models (LLMs) with few-shot prompting.
method Theoretical analysis and sensitivity experiments on transformer behavior with coherent reasoning and corrupted demonstrations.
result The transformer gains better error correction ability and more accurate predictions when coherent reasoning is integrated.

Social media sources can provide crucial information in crisis situations, but discovering relevant messages is not trivial. Methods have so far focused on universal detection models for all kinds of crises or for certain crisis types (e.g. floods). Event-specific models could implement a more focused search area, but …

2019-10-05abs ↗pdf ↗

AdaDPSyn generates synthetic examples to protect private data in ICL.

problem Protecting private data in in-context learning with large language models.
method Data-adaptive differentially private algorithm that dynamically adjusts noise level based on data properties.
result AdaDPSyn outperforms existing methods in preserving high ICL accuracy while maintaining differential privacy.

Fine-tuning harms in-context learning, but restricting updates to the value matrix improves zero-shot performance.

problem Fine-tuning harms in-context learning, reducing zero-shot performance on unseen tasks.
method Theoretical analysis of linear attention models, identifying conditions for degraded few-shot performance.
result Restricting updates to the value matrix improves zero-shot performance while preserving in-context learning.

ICON-OCnet solves optimal execution problems with neural networks and few examples.

problem Optimal order execution in markets with unknown price impact.
method Transformer-based neural network architecture (ICON-OCnet) that learns price impact from few examples and applies it to optimal execution strategies.
result ICON-OCnet accurately infers price impact models and retrieves optimal execution strategies for various propagator kernels.

Meta-learning improves GNN initializations for low-resource drug discovery.

problem Limited labeled data hinders deep learning in drug discovery.
method Model-Agnostic Meta-Learning (MAML) and its variants for graph neural networks initializations.
result Meta-initializations outperform multi-task pre-training baselines on 16 out of 20 tasks and all out-of-distribution tasks.

Study on how attention in prompt-tuning affects large language models.

problem Limited theoretical understanding of prompt-tuning and attention in LLMs.
method Exploration of prompt-tuning for one-layer attention architectures, contextual mixture-models, and self-contained prompt-attention model.
result Softmax-prompt-attention is more expressive than self-attention and linear-prompt-attention under contextual data model.

Understanding optimal prompts for binary sequence predictors is challenging.

problem Finding good prompts for binary sequence predictors is difficult.
method Viewing prompting as finding the best conditioning sequence on a near-optimal sequence predictor, using empirical and statistical analysis.
result Optimal prompts can be better understood given the pretraining distribution, which is not usually available.

Diffusion LLMs can efficiently generate harmful prompts for adversarial testing.

problem Generating harmful prompts for adversarial testing is resource-intensive and costly.
method Transformed adversarial prompt optimization into an efficient inference task using pretrained Diffusion LLMs.
result Only a few conditional samples are required to generate harmful prompts with high reward.

Transformer models can approximate smooth functions with prompts, enhancing LLMs' dynamic capabilities.

problem Lack of theoretical framework for prompt engineering in transformer models.
method Formal framework demonstrating transformer models can approximate ββ-times differentiable functions with prompts.
result Transformer models can approximate ββ-times differentiable functions with arbitrary precision using appropriately structured prompts.

Single-head transformers with a single self-attention layer can approximate any sequence-to-sequence function and are efficient under certain conditions.

problem Statistical and computational limits of prompt tuning for transformer-based models.
method Investigation of single-head transformers with a single self-attention layer, proving universality and efficiency under SETH.
result Existence of almost-linear time prompt tuning inference algorithms under certain conditions.

Fine-tuning a deep network trained with the standard cross-entropy loss is a strong baseline for few-shot learning. When fine-tuned transductively, this outperforms the current state-of-the-art on standard datasets such as Mini-ImageNet, Tiered-ImageNet, CIFAR-FS and FC-100 with the same hyper-parameters. The simplicit…

2019-09-06abs ↗pdf ↗

FROB model improves robustness and reliable confidence for few-shot OoD detection.

problem Challenges in few-shot classification and OoD detection due to limited samples and adversarial attacks.
method FROB model combines support boundary generation and few-shot Outlier Exposure (OE) for improved robustness and reliable confidence.
result FROB achieves generalization to unseen OoD and maintains robustness independent of few-shot number.

In the few-shot scenario, a learner must effectively generalize to unseen classes given a small support set of labeled examples. While a relatively large amount of research has gone into few-shot learning for image classification, little work has been done on few-shot video classification. In this work, we address the …

2019-09-14abs ↗pdf ↗

Accurate image classification given small amounts of labelled data (few-shot classification) remains an open problem in computer vision. In this work we examine how the known texture bias of Convolutional Neural Networks (CNNs) affects few-shot classification performance. Although texture bias can help in standard imag…

2019-10-18abs ↗pdf ↗

Reward collapse occurs when ranking-based reward models yield uniform rewards for different prompts.

problem Reward collapse in aligning large language models with human preferences.
method Introduced a prompt-aware optimization scheme to derive closed-form expressions for reward distributions.
result Our prompt-aware utility functions significantly alleviate reward collapse during training.

Advances few-shot classification by treating it as supervised learning and proposing new training techniques.

problem Formulating the ability of humans to learn from limited data in machine learning.
method Formulated few-shot classification as a supervised learning problem and introduced multi-episode and cross-way training techniques.
result Proposed training strategies accelerate the training process without accuracy loss.

Paper accelerates Bayesian few-shot classification using mirror descent.

problem Non-conjugate inference in Bayesian few-shot classification.
method Integrates mirror descent-based variational inference into Gaussian process-based few-shot classification.
result Accelerated convergence and improved uncertainty quantification.

eva method improves RL models by adaptively creating prompts, boosting performance.

problem Fixed prompt distribution limits scalability of RL models post-training.
method Evolving Alignment via Asymmetric Self-Play (eva) approach.
result Significant performance boost on benchmarks, e.g. 51.6% to 60.1% win-rate.

The paper studies estimation rates for MoE models with a new prompt.

problem Estimating parameters in a softmax-contaminated MoE model.
method Analytic notion of distinguishability, minimax optimal estimation rates.
result Estimation rates are minimax optimal under distinguishability, but slower otherwise.

Paper introduces negative margin loss for better few-shot classification accuracy.

problem Improving few-shot classification accuracy with metric learning.
method Introduces negative margin loss and analyzes its impact on feature discriminability.
result Negative margin loss outperforms regular softmax loss on few-shot classification benchmarks.

LanguaShrink compresses prompts using psycholinguistic principles to reduce costs.

problem Reducing computational cost and efficiency in large language model inference.
method Leverages psycholinguistic principles and the Ebbinghaus memory curve to compress prompts.
result Achieves up to 26 times compression while maintaining semantic similarity.

CosML combines domain-specific meta-learners for cross-domain few-shot classification.

problem Generalizing to unseen domains while meta-learning on multiple seen domains.
method CosML trains domain-specific meta-learners and combines their meta-parameters in the parameter space.
result CosML outperforms state-of-the-art methods and achieves strong cross-domain generalization.

VERA uses variational inference to jailbreak LLMs without manual optimization.

problem Lack of principled objective for gradient-based optimization in jailbreaking LLMs.
method VERA casts black-box jailbreak prompting as a variational inference problem, training a small attacker LLM to approximate the target LLM's posterior over adversarial prompts.
result VERA achieves strong performance across various target LLMs, demonstrating the value of probabilistic inference for adversarial prompt generation.

Paper improves few-shot classification accuracy using feature distribution preprocessing.

problem Challenges of few-shot classification due to limited labelled samples.
method Proposes a novel transfer-based method that preprocesses feature vectors to Gaussian-like distributions and uses optimal-transport inspired algorithms.
result Achieves state-of-the-art accuracy on standardized vision benchmarks.