Update README.md

README.md

@@ -53,14 +53,10 @@ tags:
 
 Boo bench is built to capture the pages that scare VLMs (cause them to make mistakes) but most specialized document OCR APIs perform well on. It is a playful and simple litmus test for understanding the need for specialized models, tools, and agent harnesses for document processing, beyond simply straight shotting documents to VLMs. 
 
-Some pages make a vision model lose its nerve: a dot leader that never ends, a grid of 192 ghosts, a code written in
-O and 0, a script it has barely seen. The frightened model keeps going, one more dot, one more row, one more ghost,
-and writes more than the page holds. **That is a Boo.**
+A **Boo** is an element on a page that scares a parser. We frequently see failure modes where certain pages cause parsers
+to go into repitition loops or stop parsing properly.
 
-**Boo Bench** is a haunted house of **179 pages built to scare document parsers into repeating themselves**, and one
-rule per page that catches it: *the output may repeat nothing more than the page holds.* Writing a leader in another
-character, splitting, merging or dropping a repetition never fails; only overshoot does. An empty answer fails too, so
-a parser cannot pass by saying nothing.
+**Boo Bench** is a haunted house of **179 pages built to scare document parsers into repeating themselves**.
 
 ## The four rooms
 
@@ -68,10 +64,10 @@ a parser cannot pass by saying nothing.
 
 | Room (split) | Pages | Ground truth | What is inside |
 |---|---:|---|---|
-| 🕯️ `real` | 100 | the page's text layer | Real pages with a repetition: dot leaders, fill-in blanks, rules of `=`, rows of `0.00`, repeated lines, in 23 kinds. Chosen from 131k candidate pages (out of 2.64M) as the ones that came closest to tricking production parsers, one per document. |
+| 🕯️ `real` | 100 | the page's text layer | Real pages with a repetition: dot leaders, fill-in blanks, rules of `=`, rows of `0.00`, repeated lines, in 23 kinds. |
 | 🕸️ `trap` | 48 | exact, generated | 16 trap families × 3 Halloween-themed pages: near-identical rows, sequences and lists that stop early, fill-in forms, repeated blocks, repetition written as an instruction (`la la la (×3)`), dense number grids, codes (Morse, binary, bar glyphs), box drawing and text art, tiled watermarks, homoglyphs (`O0`, `l1I`), stacked diacritics, right-to-left and vertical text, rare emoji sequences, nearly blank pages, giant words. |
-| 👻 `ghost` | 16 | exact, generated | Grids, rows and tallies of ghost, pumpkin, bat, spider and skull emoji: the purest temptation to loop. |
-| 💀 `script` | 15 | none needed | Scans of the UDHR's Article 1, the alphabet and the digits in 15 scripts (Armenian, Canadian syllabics, Cherokee, Ethiopic, Gujarati, Kannada, Khmer, Malayalam, Mongolian, Myanmar, Sinhala, Syriac, Tamil, Thaana, Vietnamese). The page repeats nothing, so any repetition is a loop. |
+| 👻 `ghost` | 16 | exact, generated | Grids, rows and tallies of ghost, pumpkin, bat, spider and skull emoji. |
+| 💀 `script` | 15 | none needed | Scans of the UDHR's Article 1, the alphabet and the digits in 15 scripts (Armenian, Canadian syllabics, Cherokee, Ethiopic, Gujarati, Kannada, Khmer, Malayalam, Mongolian, Myanmar, Sinhala, Syriac, Tamil, Thaana, Vietnamese). |
 
 Trap, ghost and script pages are image-only PDFs, so a parser has to read the pixels.
 
@@ -110,8 +106,7 @@ parse-bench run llamaparse_agentic --input_dir ./my_docs
 ## The rule: `repetition_overshoot`
 
 The output is cleaned of markdown and HTML syntax (table separators, grids, rules, bold markers, tags), look-alike
-characters are folded (`O`/`0`, `I`/`l`/`ǀ`/`1`, `$`/`5`) so a misread is not a loop, and the page fails on the worst
-overshoot of:
+characters are folded (`O`/`0`, `I`/`l`/`ǀ`/`1`, `$`/`5`), and the page fails on the worst overshoot of:
 
 | Budget | Limit |
 |---|---|