# Federated Instruction Tuning (SFT) on Alpaca GPT4 (held-out) Source: [Federated Instruction Tuning (SFT) on Alpaca GPT4 (held-out)]() Collection index: [llms.txt]() Data updated: 2026-07-07T09:16:36.078Z Evaluation of model performance using Federated Instruction Tuning (Supervised Fine-tuning) methodology on the held-out portion of the Alpaca GPT4 dataset, reporting loss and perplexity metrics. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Federated Instruction Tuning (Supervised Fine-tuning) | | Dataset (including any stated split) | Alpaca GPT4 (held-out) | | Metric | Loss | | Unit | SCALAR | | Direction | Lower is better | | Results loaded | 8 of 8 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Loss](): Lower is better; unit: SCALAR. - [Perplexity (PPL)](): Lower is better; unit: SCALAR. - [Token Accuracy](): Higher is better; unit: PERCENT. - [Response NLL](): Lower is better; unit: SCALAR. ## Recorded results | Method | Loss | Result date | Report paper | | --- | --- | --- | --- | | FedACT | 0.8979 | Not recorded | 6a4c880785dba69fef8b267a | | FedAvg | 0.8997 | Not recorded | 6a4c880785dba69fef8b267a | | FedProx | 0.9002 | Not recorded | 6a4c880785dba69fef8b267a | | SCAFFOLD | 0.9007 | Not recorded | 6a4c880785dba69fef8b267a | | FedAdam | 0.9046 | Not recorded | 6a4c880785dba69fef8b267a | | FedAdamW | 0.9047 | Not recorded | 6a4c880785dba69fef8b267a | | FedLADA | 0.9112 | Not recorded | 6a4c880785dba69fef8b267a | | Base | 1.0682 | Not recorded | 6a4c880785dba69fef8b267a | ### FedACT Loss: 0.8979. Result ID: 6a4cc3db60b88bb339c95f72. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Communication Rounds":"200","Number of Clients":"20","Participation Rate":"10%","Batch Size":"16","K (Local Iterations)":"10"}. This record does not assert independent reproduction. ### FedAvg Loss: 0.8997. Result ID: 6a4cc3db60b88bb339c95f6c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Communication Rounds":"200","Number of Clients":"20","Participation Rate":"10%","Batch Size":"16","K (Local Iterations)":"10"}. This record does not assert independent reproduction. ### FedProx Loss: 0.9002. Result ID: 6a4cc3db60b88bb339c95f6d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Communication Rounds":"200","Number of Clients":"20","Participation Rate":"10%","Batch Size":"16","K (Local Iterations)":"10"}. This record does not assert independent reproduction. ### SCAFFOLD Loss: 0.9007. Result ID: 6a4cc3db60b88bb339c95f6e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Communication Rounds":"200","Number of Clients":"20","Participation Rate":"10%","Batch Size":"16","K (Local Iterations)":"10"}. This record does not assert independent reproduction. ### FedAdam Loss: 0.9046. Result ID: 6a4cc3db60b88bb339c95f6f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Communication Rounds":"200","Number of Clients":"20","Participation Rate":"10%","Batch Size":"16","K (Local Iterations)":"10"}. This record does not assert independent reproduction. ### FedAdamW Loss: 0.9047. Result ID: 6a4cc3db60b88bb339c95f71. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Communication Rounds":"200","Number of Clients":"20","Participation Rate":"10%","Batch Size":"16","K (Local Iterations)":"10"}. This record does not assert independent reproduction. ### FedLADA Loss: 0.9112. Result ID: 6a4cc3db60b88bb339c95f70. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Communication Rounds":"200","Number of Clients":"20","Participation Rate":"10%","Batch Size":"16","K (Local Iterations)":"10"}. This record does not assert independent reproduction. ### Base Loss: 1.0682. Result ID: 6a4cc3db60b88bb339c95f6b. [Reporting paper]() Recorded evaluation configuration: {"Communication Rounds":"200","Number of Clients":"20","Participation Rate":"10%","Batch Size":"16","K (Local Iterations)":"10"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.