I am currently planning to finetune a GPT-4 model using my own dataset and am curious to know if there is a possibility to access the logits as well during this process. Could you provide information on whether this feature is available and how I might go about obtaining these logits?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Is it possible to get the entire logits/softmax probability vector of the output? | 5 | 8066 | December 21, 2023 | |
| New Fine-Tunning logprobs | 1 | 974 | September 1, 2023 | |
| Logit_bias for finetuned model | 0 | 141 | May 2, 2024 | |
| Will we be allowed to fine-tune o1 models in the future? | 2 | 1326 | October 6, 2024 | |
| Is it possible to use logprobs with assistants api? | 3 | 277 | August 22, 2024 |