Repository navigation
question #6382
Answered
by
pablodiazjorge
mondaljagannath161-source
asked this question in
Q&A
question
#6382
|
Can we use rea in reverse engineering a ai model or llm I mean cloning a ai exactly ? |
Answered by
pablodiazjorge
Oct 23, 2026
Replies: 1 comment
|
Not in that sense. REA analyses artifacts you already have locally, so it can't reproduce a hosted model such as ChatGPT or Claude: their weights never leave the provider. For model files you do have, main (not yet in a release; the latest published version is 6.7.0) adds |
0 replies
Answer selected by
mondaljagannath161-source
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Not in that sense. REA analyses artifacts you already have locally, so it can't reproduce a hosted model such as ChatGPT or Claude: their weights never leave the provider.
For model files you do have, main (not yet in a release; the latest published version is 6.7.0) adds
inspect_model_artifact(CLI:rea inspect-model-artifact). It reads what a local GGUF or safetensors file declares: metadata, the tensor directory and a header digest. It does not load the model, interpret the weights or run inference (docs/model-artifacts.md). So REA can tell you what a model file contains, but it can't recreate or copy a model.