Scrub README: remove paper-identifying references

README.md

@@ -27,14 +27,6 @@ configs:
 
 Hand-authored functional-correctness reference set for arith, linalg+memref, and stablehlo (n=30, 10 per dialect).
 
-This dataset is one of six NL→MLIR benchmarks released alongside the NeurIPS
-2026 Evaluations & Datasets track paper *Cross-Dialect Generalization Without
-Retraining: Benchmarks and Evaluation of Schema-Derived Constrained Decoding
-for MLIR* (anonymous submission). The full suite — `MLIR-Spec-150`,
-`Linalg-Spec-30`, `StableHLO-Spec-30`, `StableHLO-Held-Out-200`,
-`StableHLO-OutOfGrammar-25`, and `MLIR-Functional-Reference-30` — totals 465
-instances across three MLIR dialects.
-
 ## Composition
 
 - **Instances**: 30
@@ -56,41 +48,19 @@ pass-rate under the dialect's verifier is the primary evaluation metric.
 
 ## Source format
 
-For paper reproducibility, individual per-record JSON files (the
-`examples/*.json` layout used by the companion code repository) and the
-MLCommons Croissant 1.0 metadata (`croissant.json`) ship together with the
-release. The JSONL file at `data/test.jsonl` is the canonical HuggingFace
-interface; it is generated 1-to-1 from the source records.
+The JSONL file at `data/test.jsonl` is the canonical HuggingFace interface.
+MLCommons Croissant 1.0 metadata (`croissant.json`) ships alongside the
+release.
 
 ## Datasheet
 
-A full Gebru-style datasheet covering motivation, collection, preprocessing,
-uses, distribution, and maintenance is included in the companion
-reproducibility archive (`docs/datasheets/datasheet.md`). Key points:
+Key points (full Gebru-style datasheet ships with the dataset archive):
 
 - All reference MLIR programs are verifier-clean at the time of release.
-- Hand-authored single-author (no crowdsourcing, no LLM-authored references).
+- Hand-authored (no crowdsourcing, no LLM-authored references).
 - Test-only — fine-tuning on these benchmarks contaminates future evaluation
   and is explicitly out of scope.
 
-## Companion artifacts
-
-- Reproducibility archive (code + scripts): `submission_artifact.tar.gz`
-  in the OpenReview attachment / Zenodo mirror.
-- Companion code repository: <will be populated at camera-ready>.
-
-## Citation
-
-```
-@inproceedings{anonymous2026crossdialect,
-  title     = {Cross-Dialect Generalization Without Retraining: Benchmarks and Evaluation of Schema-Derived Constrained Decoding for MLIR},
-  author    = {Anonymous},
-  booktitle = {Advances in Neural Information Processing Systems (NeurIPS), Datasets and Benchmarks Track},
-  year      = {2026},
-  note      = {Anonymous submission under review.}
-}
-```
-
 ## License
 
 Apache-2.0. See `LICENSE`.