Fadhel12 commited on
Commit
ffc4538
·
verified ·
1 Parent(s): 574199c

Recover cached SimRAG paraphrases with minimal selection

Browse files
README.md CHANGED
@@ -29,15 +29,15 @@ configs:
29
  path: retrieved_answer_judged/multi_source.jsonl
30
  ---
31
 
32
- # SimRAG Tahap II — balanced cached paraphrase selection
33
 
34
  QA dan konteks retrieval berasal dari run B. Keanggotaan dua dataset dipertahankan.
35
- Seleksi parafrase: `balanced`. Keluaran mentah dipulihkan memakai hash prompt yang sama.
36
  Tidak memaksakan anti-salin, metadata, overlap kata, letak span di kalimat pertama, atau penghapusan kalimat pendukung.
37
  `response` adalah jawaban awal. `response_paraphrase` adalah target jawaban hasil pemulihan.
38
  Untuk training gunakan `prompt_paraphrase` (MS: soal gabungan alami; jenis lain: sama dengan prompt),
39
  `response_paraphrase`, `specific_instruction`, dan `context`.
40
- Seleksi berbasis aturan bukan validasi semantik semua fakta. Record tanpa kandidat yang lolos memakai fallback B.
41
  Audit pemulihan disimpan lokal agar tidak menjadi split dataset tambahan.
42
 
43
  Jumlah per config:
 
29
  path: retrieved_answer_judged/multi_source.jsonl
30
  ---
31
 
32
+ # SimRAG Tahap II — relaxed paraphrase recovery
33
 
34
  QA dan konteks retrieval berasal dari run B. Keanggotaan dua dataset dipertahankan.
35
+ Seleksi parafrase: `minimal`. Keluaran mentah dipulihkan memakai hash prompt yang sama.
36
  Tidak memaksakan anti-salin, metadata, overlap kata, letak span di kalimat pertama, atau penghapusan kalimat pendukung.
37
  `response` adalah jawaban awal. `response_paraphrase` adalah target jawaban hasil pemulihan.
38
  Untuk training gunakan `prompt_paraphrase` (MS: soal gabungan alami; jenis lain: sama dengan prompt),
39
  `response_paraphrase`, `specific_instruction`, dan `context`.
40
+ Seleksi minimal bukan validasi semantik semua fakta. Record tanpa kandidat yang lolos tetap memakai jawaban sebelumnya.
41
  Audit pemulihan disimpan lokal agar tidak menjadi split dataset tambahan.
42
 
43
  Jumlah per config:
retrieved_answer_judged/multi_source.jsonl CHANGED
The diff for this file is too large to render. See raw diff
 
retrieved_answer_judged/multiple_choice.jsonl CHANGED
The diff for this file is too large to render. See raw diff
 
retrieved_answer_judged/short_span.jsonl CHANGED
The diff for this file is too large to render. See raw diff
 
retrieved_answer_judged/unanswerable.jsonl CHANGED
The diff for this file is too large to render. See raw diff
 
retrieved_selection/multi_source.jsonl CHANGED
The diff for this file is too large to render. See raw diff
 
retrieved_selection/multiple_choice.jsonl CHANGED
The diff for this file is too large to render. See raw diff
 
retrieved_selection/short_span.jsonl CHANGED
The diff for this file is too large to render. See raw diff
 
retrieved_selection/unanswerable.jsonl CHANGED
The diff for this file is too large to render. See raw diff