PyTorchKR Papers

“Code Generation”

총 10,163편
태그·기간 등 상세 필터

카테고리 탐색

arXiv:2112.09737공식 코드 · Papers with Code

Learning to Repair: Repairing model output errors after deployment using a dynamic memory of feedback

Large language models (LMs), while powerful, are not immune to mistakes, but can be difficult to retrain. Our goal is for an LM to continue to improve after deployment, without retraining, using feedback from the user. Our approach pairs an LM with (i) a growing memory of cases where the user identified an output error and provided general feedback on how to correct it (ii) a corrector model, trained to translate this general feedback into specific edits to repair the model output. Given a new, unseen input, our model can then use feedback from similar, past cases to repair output errors that may occur. We instantiate our approach using an existing, fixed model for script generation, that takes a goal (e.g., "bake a cake") and generates a partially ordered sequence of actions to achieve that goal, sometimes containing errors. Our memory-enhanced system, FBNet, learns to apply user feedback to repair such errors (up to 30 points improvement), while making a start at avoiding similar past mistakes on new, unseen examples (up to 7 points improvement in a controlled setting). This is a first step towards strengthening deployed models, potentially broadening their utility. Our code and data is available at https://github.com/allenai/interscript/.

arXiv:2203.11006공식 코드 · Papers with Code

Underwater Light Field Retention : Neural Rendering for Underwater Imaging

Underwater Image Rendering aims to generate a true-tolife underwater image from a given clean one, which could be applied to various practical applications such as underwater image enhancement, camera filter, and virtual gaming. We explore two less-touched but challenging problems in underwater image rendering, namely, i) how to render diverse underwater scenes by a single neural network? ii) how to adaptively learn the underwater light fields from natural exemplars, i,e., realistic underwater images? To this end, we propose a neural rendering method for underwater imaging, dubbed UWNR (Underwater Neural Rendering). Specifically, UWNR is a data-driven neural network that implicitly learns the natural degenerated model from authentic underwater images, avoiding introducing erroneous biases by hand-craft imaging models. Compared with existing underwater image generation methods, UWNR utilizes the natural light field to simulate the main characteristics ofthe underwater scene. Thus, it is able to synthesize a wide variety ofunderwater images from one clean image with various realistic underwater images. Extensive experiments demonstrate that our approach achieves better visual effects and quantitative metrics over previous methods. Moreover, we adopt UWNR to build an open Large Neural Rendering Underwater Dataset containing various types of water quality, dubbed LNRUD. The source code and LNRUD are available at https: //github.com/Ephemeral182/UWNR.

arXiv:2204.00987공식 코드 · Papers with Code

BinsFormer: Revisiting Adaptive Bins for Monocular Depth Estimation

Monocular depth estimation is a fundamental task in computer vision and has drawn increasing attention. Recently, some methods reformulate it as a classification-regression task to boost the model performance, where continuous depth is estimated via a linear combination of predicted probability distributions and discrete bins. In this paper, we present a novel framework called BinsFormer, tailored for the classification-regression-based depth estimation. It mainly focuses on two crucial components in the specific task: 1) proper generation of adaptive bins and 2) sufficient interaction between probability distribution and bins predictions. To specify, we employ the Transformer decoder to generate bins, novelly viewing it as a direct set-to-set prediction problem. We further integrate a multi-scale decoder structure to achieve a comprehensive understanding of spatial geometry information and estimate depth maps in a coarse-to-fine manner. Moreover, an extra scene understanding query is proposed to improve the estimation accuracy, which turns out that models can implicitly learn useful information from an auxiliary environment classification task. Extensive experiments on the KITTI, NYU, and SUN RGB-D datasets demonstrate that BinsFormer surpasses state-of-the-art monocular depth estimation methods with prominent margins. Code and pretrained models will be made publicly available at \url{https://github.com/zhyever/Monocular-Depth-Estimation-Toolbox}.

arXiv:2205.12502공식 코드 · Papers with Code

The Dialog Must Go On: Improving Visual Dialog via Generative Self-Training

Visual dialog (VisDial) is a task of answering a sequence of questions grounded in an image, using the dialog history as context. Prior work has trained the dialog agents solely on VisDial data via supervised learning or leveraged pre-training on related vision-and-language datasets. This paper presents a semi-supervised learning approach for visually-grounded dialog, called Generative Self-Training (GST), to leverage unlabeled images on the Web. Specifically, GST first retrieves in-domain images through out-of-distribution detection and generates synthetic dialogs regarding the images via multimodal conditional text generation. GST then trains a dialog agent on the synthetic and the original VisDial data. As a result, GST scales the amount of training data up to an order of magnitude that of VisDial (1.2M to 12.9M QA data). For robust training of the synthetic dialogs, we also propose perplexity-based data selection and multimodal consistency regularization. Evaluation on VisDial v1.0 and v0.9 datasets shows that GST achieves new state-of-the-art results on both datasets. We further observe the robustness of GST against both visual and textual adversarial attacks. Finally, GST yields strong performance gains in the low-data regime. Code is available at https://github.com/gicheonkang/gst-visdial.

arXiv:2310.15444공식 코드 · Papers with Code

Fast Propagation is Better: Accelerating Single-Step Adversarial Training via Sampling Subnetworks

Adversarial training has shown promise in building robust models against adversarial examples. A major drawback of adversarial training is the computational overhead introduced by the generation of adversarial examples. To overcome this limitation, adversarial training based on single-step attacks has been explored. Previous work improves the single-step adversarial training from different perspectives, e.g., sample initialization, loss regularization, and training strategy. Almost all of them treat the underlying model as a black box. In this work, we propose to exploit the interior building blocks of the model to improve efficiency. Specifically, we propose to dynamically sample lightweight subnetworks as a surrogate model during training. By doing this, both the forward and backward passes can be accelerated for efficient adversarial training. Besides, we provide theoretical analysis to show the model robustness can be improved by the single-step adversarial training with sampled subnetworks. Furthermore, we propose a novel sampling strategy where the sampling varies from layer to layer and from iteration to iteration. Compared with previous methods, our method not only reduces the training cost but also achieves better model robustness. Evaluations on a series of popular datasets demonstrate the effectiveness of the proposed FB-Better. Our code has been released at https://github.com/jiaxiaojunQAQ/FP-Better.

arXiv:2310.17245공식 코드 · Papers with Code

CROP: Conservative Reward for Model-based Offline Policy Optimization

Offline reinforcement learning (RL) aims to optimize policy using collected data without online interactions. Model-based approaches are particularly appealing for addressing offline RL challenges due to their capability to mitigate the limitations of offline data through data generation using models. Prior research has demonstrated that introducing conservatism into the model or Q-function during policy optimization can effectively alleviate the prevalent distribution drift problem in offline RL. However, the investigation into the impacts of conservatism in reward estimation is still lacking. This paper proposes a novel model-based offline RL algorithm, Conservative Reward for model-based Offline Policy optimization (CROP), which conservatively estimates the reward in model training. To achieve a conservative reward estimation, CROP simultaneously minimizes the estimation error and the reward of random actions. Theoretical analysis shows that this conservative reward mechanism leads to a conservative policy evaluation and helps mitigate distribution drift. Experiments on D4RL benchmarks showcase that the performance of CROP is comparable to the state-of-the-art baselines. Notably, CROP establishes an innovative connection between offline and online RL, highlighting that offline RL problems can be tackled by adopting online RL techniques to the empirical Markov decision process trained with a conservative reward. The source code is available with https://github.com/G0K0URURI/CROP.git.

arXiv:2312.06988공식 코드 · Papers with Code

MWSIS: Multimodal Weakly Supervised Instance Segmentation with 2D Box Annotations for Autonomous Driving

Instance segmentation is a fundamental research in computer vision, especially in autonomous driving. However, manual mask annotation for instance segmentation is quite time-consuming and costly. To address this problem, some prior works attempt to apply weakly supervised manner by exploring 2D or 3D boxes. However, no one has ever successfully segmented 2D and 3D instances simultaneously by only using 2D box annotations, which could further reduce the annotation cost by an order of magnitude. Thus, we propose a novel framework called Multimodal Weakly Supervised Instance Segmentation (MWSIS), which incorporates various fine-grained label generation and correction modules for both 2D and 3D modalities to improve the quality of pseudo labels, along with a new multimodal cross-supervision approach, named Consistency Sparse Cross-modal Supervision (CSCS), to reduce the inconsistency of multimodal predictions by response distillation. Particularly, transferring the 3D backbone to downstream tasks not only improves the performance of the 3D detectors, but also outperforms fully supervised instance segmentation with only 5% fully supervised annotations. On the Waymo dataset, the proposed framework demonstrates significant improvements over the baseline, especially achieving 2.59% mAP and 12.75% mAP increases for 2D and 3D instance segmentation tasks, respectively. The code is available at https://github.com/jiangxb98/mwsis-plugin.

arXiv:2312.08009공식 코드 · Papers with Code

Semi-Supervised Class-Agnostic Motion Prediction with Pseudo Label Regeneration and BEVMix

Class-agnostic motion prediction methods aim to comprehend motion within open-world scenarios, holding significance for autonomous driving systems. However, training a high-performance model in a fully-supervised manner always requires substantial amounts of manually annotated data, which can be both expensive and time-consuming to obtain. To address this challenge, our study explores the potential of semi-supervised learning (SSL) for class-agnostic motion prediction. Our SSL framework adopts a consistency-based self-training paradigm, enabling the model to learn from unlabeled data by generating pseudo labels through test-time inference. To improve the quality of pseudo labels, we propose a novel motion selection and re-generation module. This module effectively selects reliable pseudo labels and re-generates unreliable ones. Furthermore, we propose two data augmentation strategies: temporal sampling and BEVMix. These strategies facilitate consistency regularization in SSL. Experiments conducted on nuScenes demonstrate that our SSL method can surpass the self-supervised approach by a large margin by utilizing only a tiny fraction of labeled data. Furthermore, our method exhibits comparable performance to weakly and some fully supervised methods. These results highlight the ability of our method to strike a favorable balance between annotation costs and performance. Code will be available at https://github.com/kwwcv/SSMP.

arXiv:2401.05430공식 코드 · Papers with Code

Multi-relational Graph Diffusion Neural Network with Parallel Retention for Stock Trends Classification

Stock trend classification remains a fundamental yet challenging task, owing to the intricate time-evolving dynamics between and within stocks. To tackle these two challenges, we propose a graph-based representation learning approach aimed at predicting the future movements of multiple stocks. Initially, we model the complex time-varying relationships between stocks by generating dynamic multi-relational stock graphs. This is achieved through a novel edge generation algorithm that leverages information entropy and signal energy to quantify the intensity and directionality of inter-stock relations on each trading day. Then, we further refine these initial graphs through a stochastic multi-relational diffusion process, adaptively learning task-optimal edges. Subsequently, we implement a decoupled representation learning scheme with parallel retention to obtain the final graph representation. This strategy better captures the unique temporal features within individual stocks while also capturing the overall structure of the stock graph. Comprehensive experiments conducted on real-world datasets from two US markets (NASDAQ and NYSE) and one Chinese market (Shanghai Stock Exchange: SSE) validate the effectiveness of our method. Our approach consistently outperforms state-of-the-art baselines in forecasting next trading day stock trends across three test periods spanning seven years. Datasets and code have been released (https://github.com/pixelhero98/MGDPR).

arXiv:2402.03201공식 코드 · Papers with Code

Guidance with Spherical Gaussian Constraint for Conditional Diffusion

Recent advances in diffusion models attempt to handle conditional generative tasks by utilizing a differentiable loss function for guidance without the need for additional training. While these methods achieved certain success, they often compromise on sample quality and require small guidance step sizes, leading to longer sampling processes. This paper reveals that the fundamental issue lies in the manifold deviation during the sampling process when loss guidance is employed. We theoretically show the existence of manifold deviation by establishing a certain lower bound for the estimation error of the loss guidance. To mitigate this problem, we propose Diffusion with Spherical Gaussian constraint (DSG), drawing inspiration from the concentration phenomenon in high-dimensional Gaussian distributions. DSG effectively constrains the guidance step within the intermediate data manifold through optimization and enables the use of larger guidance steps. Furthermore, we present a closed-form solution for DSG denoising with the Spherical Gaussian constraint. Notably, DSG can seamlessly integrate as a plugin module within existing training-free conditional diffusion methods. Implementing DSG merely involves a few lines of additional code with almost no extra computational overhead, yet it leads to significant performance improvements. Comprehensive experimental results in various conditional generation tasks validate the superiority and adaptability of DSG in terms of both sample quality and time efficiency.

arXiv:2406.09056공식 코드 · Papers with Code

Towards Reliable Detection of LLM-Generated Texts: A Comprehensive Evaluation Framework with CUDRT

The increasing prevalence of large language models (LLMs) has significantly advanced text generation, but the human-like quality of LLM outputs presents major challenges in reliably distinguishing between human-authored and LLM-generated texts. Existing detection benchmarks are constrained by their reliance on static datasets, scenario-specific tasks (e.g., question answering and text refinement), and a primary focus on English, overlooking the diverse linguistic and operational subtleties of LLMs. To address these gaps, we propose CUDRT, a comprehensive evaluation framework and bilingual benchmark in Chinese and English, categorizing LLM activities into five key operations: Create, Update, Delete, Rewrite, and Translate. CUDRT provides extensive datasets tailored to each operation, featuring outputs from state-of-the-art LLMs to assess the reliability of LLM-generated text detectors. This framework supports scalable, reproducible experiments and enables in-depth analysis of how operational diversity, multilingual training sets, and LLM architectures influence detection performance. Our extensive experiments demonstrate the framework's capacity to optimize detection systems, providing critical insights to enhance reliability, cross-linguistic adaptability, and detection accuracy. By advancing robust methodologies for identifying LLM-generated texts, this work contributes to the development of intelligent systems capable of meeting real-world multilingual detection challenges. Source code and dataset are available at GitHub.

arXiv:2406.13439공식 코드 · Papers with Code

Finding Blind Spots in Evaluator LLMs with Interpretable Checklists

Large Language Models (LLMs) are increasingly relied upon to evaluate text outputs of other LLMs, thereby influencing leaderboards and development decisions. However, concerns persist over the accuracy of these assessments and the potential for misleading conclusions. In this work, we investigate the effectiveness of LLMs as evaluators for text generation tasks. We propose FBI, a novel framework designed to examine the proficiency of Evaluator LLMs in assessing four critical abilities in other LLMs: factual accuracy, instruction following, coherence in long-form writing, and reasoning proficiency. By introducing targeted perturbations in answers generated by LLMs, that clearly impact one of these key capabilities, we test whether an Evaluator LLM can detect these quality drops. By creating a total of 2400 perturbed answers covering 22 perturbation categories, we conduct a comprehensive study using different evaluation strategies on five prominent LLMs commonly used as evaluators in the literature. Our findings reveal significant shortcomings in current Evaluator LLMs, which failed to identify quality drops in over 50\% of cases on average. Single-answer and pairwise evaluations demonstrated notable limitations, whereas reference-based evaluations showed comparatively better performance. These results underscore the unreliable nature of current Evaluator LLMs and advocate for cautious implementation in practical applications. Code and data are available at https://github.com/AI4Bharat/FBI.

arXiv:2406.18916공식 코드 · Papers with Code

TrustUQA: A Trustful Framework for Unified Structured Data Question Answering

Natural language question answering (QA) over structured data sources such as tables and knowledge graphs have been widely investigated, especially with Large Language Models (LLMs) in recent years. The main solutions include question to formal query parsing and retrieval-based answer generation. However, current methods of the former often suffer from weak generalization, failing to dealing with multi-types of sources, while the later is limited in trustfulness. In this paper, we propose TrustUQA, a trustful QA framework that can simultaneously support multiple types of structured data in a unified way. To this end, it adopts an LLM-friendly and unified knowledge representation method called Condition Graph(CG), and uses an LLM and demonstration-based two-level method for CG querying. For enhancement, it is also equipped with dynamic demonstration retrieval. We have evaluated TrustUQA with 5 benchmarks covering 3 types of structured data. It outperforms 2 existing unified structured data QA methods. In comparison with the baselines that are specific to one data type, it achieves state-of-the-art on 2 of the datasets. Further more, we have demonstrated the potential of our method for more general QA tasks, QA over mixed structured data and QA across structured data. The code is available at https://github.com/zjukg/TrustUQA.

arXiv:2409.12960

LVCD: Reference-based Lineart Video Colorization with Diffusion Models

We propose the first video diffusion framework for reference-based lineart video colorization. Unlike previous works that rely solely on image generative models to colorize lineart frame by frame, our approach leverages a large-scale pretrained video diffusion model to generate colorized animation videos. This approach leads to more temporally consistent results and is better equipped to handle large motions. Firstly, we introduce Sketch-guided ControlNet which provides additional control to finetune an image-to-video diffusion model for controllable video synthesis, enabling the generation of animation videos conditioned on lineart. We then propose Reference Attention to facilitate the transfer of colors from the reference frame to other frames containing fast and expansive motions. Finally, we present a novel scheme for sequential sampling, incorporating the Overlapped Blending Module and Prev-Reference Attention, to extend the video diffusion model beyond its original fixed-length limitation for long video colorization. Both qualitative and quantitative results demonstrate that our method significantly outperforms state-of-the-art techniques in terms of frame and video quality, as well as temporal consistency. Moreover, our method is capable of generating high-quality, long temporal-consistent animation videos with large motions, which is not achievable in previous works. Our code and model are available at https://luckyhzt.github.io/lvcd.

arXiv:2411.07240공식 코드 · Papers with Code

UTMath: Math Evaluation with Unit Test via Reasoning-to-Coding Thoughts

The evaluation of mathematical reasoning capabilities is essential for advancing Artificial General Intelligence (AGI). While Large Language Models (LLMs) have shown impressive performance in solving mathematical problems, existing benchmarks such as GSM8K and MATH present limitations, including narrow problem definitions with specific numbers and reliance on predetermined rules that hinder accurate assessments of reasoning and generality. This paper introduces the UTMath Benchmark, a robust evaluation framework designed to assess LLMs through extensive unit tests, with a focus on both the accuracy and generality of model responses. It comprises 1,053 cutting-edge problems spanning nine mathematical domains, with an average of 68 test cases per problem. UTMath is highly challenging, with the best-performing model, o1-mini, solving only 32.57\% of the problems, followed by o1-preview at 27.16\%, and GPT-4o at 26.93\%. Furthermore, we present the Reasoning-to-Coding of Thoughts (RCoT) approach, which encourages LLMs to engage in explicit reasoning prior to code generation, thereby facilitating the production of more sophisticated solutions and enhancing overall performance and efficiency. Additionally, we also release the UTMath-Train training dataset (more than 70k samples), to support the community in further exploring mathematical reasoning. Our benchmark can be accessed via the following link: https://github.com/UTMathGroup/UTMath

arXiv:2412.07769공식 코드 · Papers with Code

BiMediX2: Bio-Medical EXpert LMM for Diverse Medical Modalities

This paper introduces BiMediX2, a bilingual (Arabic-English) Bio-Medical EXpert Large Multimodal Model (LMM) with a unified architecture that integrates text and visual modalities, enabling advanced image understanding and medical applications. BiMediX2 leverages the Llama3.1 architecture and integrates text and visual capabilities to facilitate seamless interactions in both English and Arabic, supporting text-based inputs and multi-turn conversations involving medical images. The model is trained on an extensive bilingual healthcare dataset consisting of 1.6M samples of diverse medical interactions for both text and image modalities, mixed in Arabic and English. We also propose the first bilingual GPT-4o based medical LMM benchmark named BiMed-MBench. BiMediX2 is benchmarked on both text-based and image-based tasks, achieving state-of-the-art performance across several medical benchmarks. It outperforms recent state-of-the-art models in medical LLM evaluation benchmarks. Our model also sets a new benchmark in multimodal medical evaluations with over 9% improvement in English and over 20% in Arabic evaluations. Additionally, it surpasses GPT-4 by around 9% in UPHILL factual accuracy evaluations and excels in various medical Visual Question Answering, Report Generation, and Report Summarization tasks. The project page including source code and the trained model, is available at https://github.com/mbzuai-oryx/BiMediX2.

arXiv:2412.13746공식 코드 · Papers with Code

RAG-RewardBench: Benchmarking Reward Models in Retrieval Augmented Generation for Preference Alignment

Despite the significant progress made by existing retrieval augmented language models (RALMs) in providing trustworthy responses and grounding in reliable sources, they often overlook effective alignment with human preferences. In the alignment process, reward models (RMs) act as a crucial proxy for human values to guide optimization. However, it remains unclear how to evaluate and select a reliable RM for preference alignment in RALMs. To this end, we propose RAG-RewardBench, the first benchmark for evaluating RMs in RAG settings. First, we design four crucial and challenging RAG-specific scenarios to assess RMs, including multi-hop reasoning, fine-grained citation, appropriate abstain, and conflict robustness. Then, we incorporate 18 RAG subsets, six retrievers, and 24 RALMs to increase the diversity of data sources. Finally, we adopt an LLM-as-a-judge approach to improve preference annotation efficiency and effectiveness, exhibiting a strong correlation with human annotations. Based on the RAG-RewardBench, we conduct a comprehensive evaluation of 45 RMs and uncover their limitations in RAG scenarios. Additionally, we also reveal that existing trained RALMs show almost no improvement in preference alignment, highlighting the need for a shift towards preference-aligned training.We release our benchmark and code publicly at https://huggingface.co/datasets/jinzhuoran/RAG-RewardBench/ for future work.

arXiv:2503.09894공식 코드 · Papers with Code

What's In Your Field? Mapping Scientific Research with Knowledge Graphs and Large Language Models

The scientific literature's exponential growth makes it increasingly challenging to navigate and synthesize knowledge across disciplines. Large language models (LLMs) are powerful tools for understanding scientific text, but they fail to capture detailed relationships across large bodies of work. Unstructured approaches, like retrieval augmented generation, can sift through such corpora to recall relevant facts; however, when millions of facts influence the answer, unstructured approaches become cost prohibitive. Structured representations offer a natural complement -- enabling systematic analysis across the whole corpus. Recent work enhances LLMs with unstructured or semistructured representations of scientific concepts; to complement this, we try extracting structured representations using LLMs. By combining LLMs' semantic understanding with a schema of scientific concepts, we prototype a system that answers precise questions about the literature as a whole. Our schema applies across scientific fields and we extract concepts from it using only 20 manually annotated abstracts. To demonstrate the system, we extract concepts from 30,000 papers on arXiv spanning astrophysics, fluid dynamics, and evolutionary biology. The resulting database highlights emerging trends and, by visualizing the knowledge graph, offers new ways to explore the ever-growing landscape of scientific knowledge. Demo: abby101/surveyor-0 on HF Spaces. Code: https://github.com/chiral-carbon/kg-for-science.

arXiv:2505.19354

GC-KBVQA: A New Four-Stage Framework for Enhancing Knowledge Based Visual Question Answering Performance

Knowledge-Based Visual Question Answering (KB-VQA) methods focus on tasks that demand reasoning with information extending beyond the explicit content depicted in the image. Early methods relied on explicit knowledge bases to provide this auxiliary information. Recent approaches leverage Large Language Models (LLMs) as implicit knowledge sources. While KB-VQA methods have demonstrated promising results, their potential remains constrained as the auxiliary text provided may not be relevant to the question context, and may also include irrelevant information that could misguide the answer predictor. We introduce a novel four-stage framework called Grounding Caption-Guided Knowledge-Based Visual Question Answering (GC-KBVQA), which enables LLMs to effectively perform zero-shot VQA tasks without the need for end-to-end multimodal training. Innovations include grounding question-aware caption generation to move beyond generic descriptions and have compact, yet detailed and context-rich information. This is combined with knowledge from external sources to create highly informative prompts for the LLM. GC-KBVQA can address a variety of VQA tasks, and does not require task-specific fine-tuning, thus reducing both costs and deployment complexity by leveraging general-purpose, pre-trained LLMs. Comparison with competing KB-VQA methods shows significantly improved performance. Our code will be made public.

arXiv:2608.09805v2cs.LG

변분 학습을 통한 RLVR 매개변수 탐험

Parameter Exploration for RLVR via Variational Learning

Exploration has been a focus of reinforcement learning research for a long time. Recently, there has been growing evidence that it is also an important ingredient in LLM reinforcement learning recipes that can significantly impact downstream performance. Many existing methods control exploration in the action-space, for example, using temperature scaling. However, these methods cannot reorder tokens but only influence the variance in the output distribution. This limits exploration and can lead to divergence or stalled training. Here, we investigate parameter-space exploration, where rollouts are generated by sampling different policies from a posterior that may each explore different rollouts. Sampling less or more diverse policies is then a complementary control lever over exploration. We introduce a family of methods called Perturbed Parameter Policy Optimization (3PO) which use different sampling strategies and different rollout grouping for reward estimation. Experiments on Olmo-3-1025-7B and Qwen2.5-Math-7B across mathematical reasoning and code generation tasks show that these approaches can improve the average downstream performance over standard GRPO at a near-identical FLOPs cost. Moreover, using multiple parameter samples generally produces fewer zero-advantage groups and malformed or incorrect rollouts during training than GRPO and action-space baselines. Overall, our work presents evidence that parameter-space exploration can improve reinforcement learning for LLMs.