Article
Fine-tuning a 350M Model for Better Structured Outputs in 100 GRPO Steps
- +1
iamleonie, burtenshaw, sergiopaniego
• • 137
🔄 In a Training Loop
Tashin Ahmed
745H1N
AI & ML interests
Neural Network Architectures
Recent Activity
upvoted an article 8 days ago
Fine-tuning a 350M Model for Better Structured Outputs in 100 GRPO Steps updated a Space about 2 months ago
745H1N/grok-grokking-repro updated a Space about 2 months ago
745H1N/repro-compositional-spectralOrganizations
upvoted an article 8 days ago
updated 19
Spaces about 2 months ago
Repro - To Grok Grokking: Provable Grokking in Ridge Regression
ðŸ§
Explore project logs and collaborate with an AI agent
Reproduction: Deep Learning of Compositional Targets with Hierarchical Spectral Methods
🎯
Explore and manage experiment logs with an interactive web logbook
Reproduction: Decision Tree Learning on Product Spaces
🎯
Explore code logs, execution traces, and workspace in one view
Reproduction: Pruning at Initialisation through the lens of Graphon Limit
🎯
Explore code logs, traces, and workspace in a web logbook
Reproduction: GradientStabilizer: Fix the Norm, Not the Gradient
🎯
Browse experiment logs, code, and traces online
Reproduction: Optimal Anytime Algorithms for Online Convex Optimization with Adversarial Constraints
🎯
Track and explore experiment logs with an interactive web logbook
Reproduction: What Does Preference Learning Recover from Pairwise Comparison Data
🎯
Explore experiment logs, traces, and workspace in a web UI
Reproduction: Rethinking Neural Network Learning Rates: A Stackelberg Perspective
🎯
Explore and manage code logs, traces, and workspace
Reproduction: High-Probability Convergence Guarantees of Decentralized SGD
🎯
Track and view experiment logs with AI collaboration
Reproduction: Near-Optimal Private Linear Regression via Iterative Hessian Mixing
🎯
Explore code logs, traces, and workspace in an interactive UI
Reproduction: MoSSP: A Momentum-Based Single-Loop Stochastic Penalty Method
🎯
Explore experiment logs, traces, and workspace online
Reproduction: Safeguarded Stochastic Polyak Step Sizes for Non-smooth Optimization
🎯
Explore experiment logs, code, and traces
Reproduction: Fairness in Aggregation: Optimal Top-k and Improved Full Ranking
🎯
Explore logs, traces, and workspace in an interactive dashboard
Reproduction: Is Spurious Correlation Removal Always Learnable
🎯
Explore code logs, traces, and workspace in a unified view
Reproduction: An Algebraic View of the Expressivity of Recurrent Language Models
🎯
Explore code logs, traces, and workspace in an interactive logbook
Reproduction: Server-Proximal Aggregation for Federated Domain-Incremental Learning
🎯
Browse project logs, traces and workspace in an interactive logbook
Reproduction: Even Faster Kernel Matrix Linear Algebra via Density Estimation
🎯
Explore and collaborate on research logbooks online
Reproduction: Provable Accuracy Collapse of Embedding-Based Representations
🎯
Browse code logs, traces, and workspace in an interactive logbook
Reproduction: Approximation Error Bounds for Holder Class with Transformers
🎯
View code logs, execution traces, and workspace together