You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Malay filenames in the question come back translated (lama.txt -> old.txt) #51
Quoting the name protects it; a bare Malay-word filename is translated to its English equivalent. Likely cause is the pool: NL was translated from English rows whose command names the file (rm old.txt), the translation API turned old.txt into lama.txt in the NL while the command stayed byte-identical, so the model learned lama.txt -> old.txt as a mapping. Filenames need to be on the stoplist the same way technical nouns are (CLAUDE.md, dataset work), or the NL side needs the original filename restored after translation.
Check: grep -c 'lama.txt' dataset/out/pool_v5.jsonl against rows whose cmd has old.txt.
README notes the quoting workaround for now. Same family as #47 (data, not code); measure with a handful of exact-output prompts, then a run.
Measured (2026-08-18), details in RESULTS.md on #67.
The hypothesis here is falsified: grep lama.txt | grep old.txt is 0 in every pool, and every NL filename absent from its command (219 rows in pool_v9) is a legitimate mention. The mapping is the base model's translation prior; what beats it is head rows carrying Malay filenames byte-identical on both sides.
Eight exact-output prompts, now EXACT in training/probe.py (#51 exact n/8):
run 10/11 models (head rows): 7/8 — only nak rename fail baru.txt jadi lama.txt → mv new.txt old.txt
run 12 (mv {t} {t2} block): 7/8 — rename right on every name pair except baru.txt, a name in zero pool rows
run 13 (baru.txt in the {t} list, 220 rows): 8/8 on SFT and DPO
Not shipped: run 13's BM ALFA is 0.423 vs 0.497, CI [−0.125, −0.022] — the first head-row-family model where the CI clears zero. Data fix is in #67 and rides with the next model that clears ALFA. README quoting workaround stays until then.
Activity
opariffazman commentedon Aug 17, 2026
Measured (2026-08-18), details in RESULTS.md on #67.
The hypothesis here is falsified:
grep lama.txt | grep old.txtis 0 in every pool, and every NL filename absent from its command (219 rows in pool_v9) is a legitimate mention. The mapping is the base model's translation prior; what beats it is head rows carrying Malay filenames byte-identical on both sides.Eight exact-output prompts, now
EXACTintraining/probe.py(#51 exact n/8):rm old.txt,cat note.txt,cp report.pdf work/, …)nak rename fail baru.txt jadi lama.txt→mv new.txt old.txtmv {t} {t2}block): 7/8 — rename right on every name pair exceptbaru.txt, a name in zero pool rowsbaru.txtin the{t}list, 220 rows): 8/8 on SFT and DPONot shipped: run 13's BM ALFA is 0.423 vs 0.497, CI [−0.125, −0.022] — the first head-row-family model where the CI clears zero. Data fix is in #67 and rides with the next model that clears ALFA. README quoting workaround stays until then.
Merge pull request #67 from officialdad/fix/51-malay-filenames