Update questions, pages, answers and document files

README.md

@@ -34,7 +34,7 @@ harnesses, and every answer, score and tool call is here. Variant: **public**.
 | --- | ---: | --- |
 | questions | 50 | Question, gold answer, scoring rule, evidence pages. Splits follow the evaluation stages. |
 | pages | 71 | Page image, converted Markdown, source URL, licence, the questions that selected the page. |
-| answers | 2845 | Model, quantization, provider, seed, configuration, repeat, answer, score, tool calls, seconds, tokens, full tool trace. |
+| answers | 3085 | Model, quantization, provider, seed, configuration, repeat, answer, score, tool calls, seconds, tokens, full tool trace. |
 
 ## Evaluation stages
 
@@ -56,6 +56,8 @@ harnesses, and every answer, score and tool call is here. Variant: **public**.
   names its `model`, `quantization`, `hosted_provider` and the `seed` sent; at temperature 0 the seed has little effect.
 - Iterations: Trace Fixes is configuration `merged_grep_revision2`, Token Reduction is `merged_grep_revision3`, and
   Thinking is `merged_grep_revision3_thinking`.
+- `merged_bm25_loop` and `merged_fusion_loop` are the grep agent with a ranked search tool, added after the test scores of the
+  other configurations were known.
 
 ## Files
 

answers/train.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:89e147f8cbd2b020ab8063f696f82a4fc8be3a9729c25eba85de09c6b3520831
-size 548851
+oid sha256:ca16d232e1e22777887ce883ebfa7b66ae66dcc4a007688b0e97f3cf96a1126c
+size 613337