agentscope-ai / agentscope-ai/Trinity-RFT

[FEATURE] Recommended way of producing multiple Experiences in multi-turn dialogs

Đang mở
#269 1 bình luận 0 reaction 0 người được giao Xem trên GitHub
enhancement
Ngôn ngữ chính
Python
Star
701
Fork
79
Merge trung bình
8 giờ 7 phút
Pull request đã merge (30 ngày)
1

Mô tả

Imagine a two-turn dialog:
1. system prompt
2. first user prompt
3. first assistant response
4. second user prompt
5. final assistant response

One could represent it as Experience by concatenating everything and masking everything except two assistant responses

Another way could be splitting it into two experiences.
The first one:
1. system prompt
2. first user prompt
3. first assistant response (masked as target assistant)

The second one:
1. system prompt
2. first user prompt
3. first assistant response
4. second user prompt
5. final assistant response (masked as target assistant)

Is this other multi-experience way recommended? If so, how to best produce it?

E.g. should a `run()` method return these several experiences, right?

Hướng dẫn đóng góp

Mở hướng dẫn đóng góp

Đánh giá

Issue này chưa được đánh giá.

Nhận issue mới trong hộp thư của bạn

Bản tóm tắt ngắn những issue GitHub phù hợp với người mới.