polished readme

README.md

@@ -54,7 +54,7 @@ configs:
 
 # CRITICS Science MT Benchmark (critics-SciMTBench)
 
-**A multilingual benchmark for evaluating machine translation of scientific text into Spanish and Basque.** Developed in the context of **CRITICS — Critical Science Without Borders**, a project on LLM-based access to scientific knowledge across linguistic boundaries.
+**A multilingual benchmark for document-level evaluation of machine translation of scientific text into Spanish and Basque.** Developed in the context of **CRITICS - Critical Science Without Borders**, a project on LLM-based access to scientific knowledge across linguistic boundaries.
 
 > **Evaluation-only release.** The benchmark must not be used for model training, fine-tuning, adaptation, or synthetic training-data generation. Access controls and the rights status of underlying source material require independent attention; see [Access, permitted use, and rights](#access-permitted-use-and-rights).
 
@@ -62,7 +62,7 @@ configs:
 |---|---|
 | Source language | English (`en`) |
 | Target languages | Spanish (`es`); Basque (`eu`) |
-| Main task | Scientific machine translation evaluation |
+| Main task | Scientific machine translation evaluation at document level|
 | Source domains | Scientific excerpts originating from arXiv and ChemRxiv |
 | Available configurations | `en-es`, `en-eu` |
 | Available splits | `test` only for each configuration |
@@ -72,21 +72,18 @@ configs:
 
 ## 1. Dataset overview
 
-**critics-SciMTBench** is a working evaluation resource for scientific machine translation. The supplied dataset-card description identifies English scientific excerpts and corresponding machine-translated texts in Spanish and Basque, with provision for human-post-edited references when independently supplied and approved. At the present stage, human references are `null`; this is **not** a reference-based benchmark release.
+**critics-SciMTBench** is a working evaluation resource for **scientific machine translation at document level**. Currently it contains English scientific excerpts and corresponding machine-translated texts in Spanish and Basque, with provision for human-post-edited references when independently supplied and approved. At the present stage, human references are `null`; this is **not** a reference-based benchmark release.
 
 The resource is intended to support research on scientific translation in multilingual contexts, particularly translation quality, terminology, and reliable assessment of scientific content. The broader CRITICS project also investigates document-level translation and multidimensional evaluation methods. These are *project research objectives*, not measured capabilities or validated results of the present dataset.
 
 ### Scope and present limitations
 
 - The published dataset configurations are **English–Spanish** and **English–Basque**; neither additional language pairs nor reference translations should be inferred from the project-level language agenda.
-- Both configurations expose a `test` split; there are no advertised `train` or `validation` partitions.
-- The source documentation describes **10 input batches** and identifiers stable at the sample and batch level; original article/document keys remain provisional until reconciled.
-- No automatic quality scores, human agreement statistics, linguistic-domain breakdowns, or benchmark results are claimed here because they are not established by the supplied documentation.
+- Both configurations expose a `test` split; there are no `train` or `validation` partitions available.
+- No automatic quality scores, human agreement statistics, linguistic-domain breakdowns, or benchmark results are coming soon.
 
 ## 2. Loading the dataset
 
-The following examples use the actual Hugging Face repository identifier and the configurations declared in this card. **Access requires repository approval and authentication if gating is enabled in the repository settings.**
-
 ```bash
 pip install -U datasets huggingface_hub
 hf auth login
@@ -112,7 +109,6 @@ To inspect the English source for the first record (using the field name describ
 ```python
 print(en_eu[0]["source_text"])
 ```
-
 **Do not combine the two language-pair configurations as if they were independent source documents.** Related records may originate from the same English excerpts; preserve sample and batch IDs when carrying out cross-language comparisons or computing uncertainty estimates.
 
 ### Obtaining supplementary files
@@ -130,7 +126,7 @@ local_repo = snapshot_download(
 print(local_repo)
 ```
 
-The directories below describe the **working repository organization** in the supplied README. Their presence in the currently published Hub repository has not been independently verified; check the repository's **Files and versions** tab before relying on these paths.
+The directories below describe the **working repository organization**:
 
 ```text
 critics_science_mt_benchmark/
@@ -141,14 +137,12 @@ critics_science_mt_benchmark/
 │   ├── pairs/
 │   │   ├── en-es/test.jsonl       # Config: en-es
 │   │   ��── en-eu/test.jsonl       # Config: en-eu
-│   ├── canonical/                # Optional; rights review required
+│   ├── canonical/                # canonical files generated from the original input (batches)
 │   └── reports/                  # Structural ingestion reports
-├── original_input/              # Original files from 10 batches; optional
-└── post_edits/                  # Optional restricted human references
+├── original_input/              # Original files from 10 batches
+└── post_edits/                  # human validated reference translations
 ```
 
-Original source files, canonical multi-language exports, and post-edited references are **not automatically included** in either `test` configuration. Do not publish rights-restricted originals or protected references merely to make the repository structure match this illustration.
-
 ## 3. Data organization and provenance
 
 The dataset is organized by **translation direction**, with two language-pair JSONL evaluation files:
@@ -158,29 +152,25 @@ The dataset is organized by **translation direction**, with two language-pair JS
 | `en-es` | English | Spanish | `test` | `data/pairs/en-es/test.jsonl` |
 | `en-eu` | English | Basque | `test` | `data/pairs/en-eu/test.jsonl` |
 
-The source material is described as coming from **arXiv and ChemRxiv**. The working pipeline distinguishes original input batches, a canonical multilingual representation, and pairwise evaluation exports. The original files are intended to be preserved unchanged; the canonical representation is an authoritative *derived* internal format and need not be published.
-
-A provisional article or document key is not a verified bibliographic identifier. Resolve article attribution and applicable licenses at the article level before any broader release of texts or translated derivatives.
+The source material is extracted from **arXiv and ChemRxiv**. The working pipeline distinguishes original input batches, a canonical multilingual representation, and pairwise evaluation exports for training and inference with language models. The original files are intended to be preserved unchanged; the canonical representation is an authoritative *derived* internal format.
 
 ### References and scoring
 
-Human-post-edited reference translations are currently **unavailable** (`null` in the supplied description). Therefore:
+Human-post-edited reference translations are currently **work in progress** (`null` in the supplied description). Therefore:
 
-- Do not treat an existing MT output as a gold-standard translation.
-- Do not report reference-based quality metrics as if verified human references were present.
-- If reference translations are incorporated later, document who produced or reviewed them, the associated rights, revision history, and the evaluation protocol.
-- Keep protected evaluation references out of training corpora, synthetic-data pipelines, and public search indexes.
+- Available MT outputs should be used as gold-standard or human-validated reference translations.
+- Reference-based quality metrics should not be reported as if verified human references were present.
 
 ## 4. CRITICS project
 
-**CRITICS — Critical Science Without Borders** (*LLMs for Translation of Scientific Knowledge in Multilingual Contexts*) is a research project funded under **CHIST-ERA IV Cofund 2025**, within the **Science in your Own Language (SOL)** topic. Its aim is to improve multilingual access to scientific knowledge while supporting scientific literacy, evidence-based argumentation, and critical thinking.
+**CRITICS - Critical Science Without Borders** (*LLMs for Translation of Scientific Knowledge in Multilingual Contexts*) is a research project funded under **CHIST-ERA IV Cofund 2025**, within the **Science in your Own Language (SOL)** topic. Its aim is to improve multilingual access to scientific knowledge while supporting scientific literacy, evidence-based argumentation, and critical thinking.
 
 | Project detail | Information |
 |---|---|
 | Project name | Critical Science Without Borders (CRITICS) |
 | Programme | CHIST-ERA IV Cofund 2025 - Science in your Own Language (SOL) |
 | Project period | 15 December 2025 - 14 December 2028 (36 months) |
-| Project coordinator| Rodrigo Agerri HiTZ Center - Ixa, University of the Basque Country EHU|
+| **Project coordinator**| **Rodrigo Agerri** HiTZ Center - Ixa, University of the Basque Country EHU|
 | Project website | [hitz.eus/critics](https://hitz.eus/critics) |
 
 ### Consortium and research responsibilities
@@ -192,24 +182,24 @@ Human-post-edited reference translations are currently **unavailable** (`null` i
 | **CNRS and UCA** | France | Argument mining and assessment (WP4) |
 | **Zurich University of Applied Sciences (ZHAW)** | Switzerland | Evaluation and applications (WP5–WP6) |
 
-The project pursues four connected objectives:
+The project states four connected objectives:
 
 1. **LLM-based translation of scientific documents**, emphasizing document-level fidelity and specialized terminology in low- and medium-resource language settings.
 2. **Argument mining and misconception detection**, focusing on evidence-based reasoning and scientific discourse.
 3. **Automatic assessment and critical-thinking support**, including critical questions, feedback, and consistent evaluation across languages.
 4. **Evaluation and benchmarking**, investigating methods beyond traditional reference-overlap metrics, including multidimensional LLM-as-a-judge methods and human comparison.
 
-**How this dataset contributes:** critics-SciMTBench provides scientific-text evaluation material for the translation and benchmarking strand of CRITICS. It is complementary to, but distinct from, the [CRITICS Science Assessments dataset](https://huggingface.co/datasets/HiTZ/critics_science_assessments), which focuses on classifying student responses against assessment rubrics. The scientific translation benchmark does **not** itself supply instructor assessments or automatically generated feedback.
+**How this dataset contributes:** critics-SciMTBench provides scientific-text evaluation material for the translation and benchmarking of MT at document level. It is complementary to, but distinct from, the [CRITICS Science Assessments dataset](https://huggingface.co/datasets/HiTZ/critics_science_assessments), which focuses on student responses with respect to data-induced assessment rubrics.
 
 For more information: [CRITICS project website](https://hitz.eus/critics) · [HiTZ Center](https://www.hitz.eus/)
 
-## 5. Intended use and methodological precautions
+## 5. Intended use
 
-**Permitted goal (subject to repository access conditions and underlying rights):** evaluate and compare scientific machine translation systems, including careful analysis of terminology, semantic preservation, and cross-language behavior.
+**Permitted goal (subject to repository access conditions and underlying rights):** evaluate and compare scientific machine translation systems at document level, including careful analysis of terminology, semantic preservation, and cross-language behavior.
 
 **Not permitted under the stated benchmark-use conditions:** use for training, fine-tuning, model adaptation, or synthetic training-data generation; republish, mirror, or transfer benchmark texts or derivatives to publicly crawlable repositories or training corpora.
 
-Evaluation users should report the language-pair configuration, dataset revision, batch/sample identifiers where relevant, model and decoding configuration, evaluation prompts or scoring protocol, and any human review procedures. Avoid presenting experimental model outputs as human reference translations.
+Evaluation users should report the language-pair configuration, dataset revision, batch/sample identifiers where relevant, model and decoding configuration, evaluation prompts or scoring protocol, and any human review procedures.
 
 ## 6. Access, permitted use, and rights
 
@@ -217,9 +207,9 @@ The benchmark is intended to be distributed as **gated, evaluation-only material
 
 **Source and derivative rights remain under review.** arXiv and ChemRxiv host materials under potentially different article-level terms. Online availability does not itself authorize redistribution, translation, or relicensing. Before releasing any scientific excerpt or derivative, check the specific work's license, permissions, and any third-party content.
 
-**License metadata note:** The supplied README contained a conflict: `license: cc-by-4.0` in YAML but prose stating that article-level redistribution licenses remain unresolved. This revised draft uses `license: other` **provisionally** to avoid asserting CC BY 4.0 for all underlying material. This is not a legal license grant. The maintainer must select and document appropriate licensing or access terms once the source-rights review is complete. The evaluation-only conditions may impose restrictions that are not compatible with an unrestricted CC BY 4.0 claim.
+**License metadata note:** Remains pending to select and document appropriate licensing or access terms once the source-rights review (arXiv and ChemRxiv) is complete.
 
-## 7. Citation and maintenance
+## 7. Citation
 
 If you use this benchmark, cite the **CRITICS project**, a published paper coming soon.