Skip to content

docs: expand disconnected curated collections guide - #139

Merged
lioraron merged 3 commits into
mainfrom
docs/rhai-16267-disconnected-curated-guide
Nov 4, 2026
Merged

lioraron merged 3 commits into
mainfrom
docs/rhai-16267-disconnected-curated-guide

Conversation

@lioraron

@lioraron lioraron commented Nov 4, 2026 •

Copy link
Copy Markdown
Contributor

Summary

  • Expand the disconnected-cluster guide with per-benchmark asset requirements and example S3 prefixes for all 24 curated benchmark entries across nine collections.
  • Document staged-data layouts and tokenizer handling for LM Evaluation Harness, LightEval, RULER, and Inspect (HLE, DocVQA, CyberSecEval 2, and BFCL), plus target/judge connectivity and adapter image verification.
  • Correct the Crows Pairs mapping to jannalu/crows_pairs_multilingual with the English subset.

Related: RHAI-16267

Validation

  • astro check: 0 errors, warnings, or hints.
  • astro build: passed (the repository reports existing component-override and bundle-size warnings).
  • Parsed all six JSON examples, confirmed 24 matrix entries, and ran git diff --check.

Summary by CodeRabbit

  • Documentation
    • Expanded the disconnected-cluster guide with asset requirements and staging instructions for 24 benchmarks across nine collections.
    • Clarified which endpoints and asset sources must be reachable during evaluation, including judge connectivity and secrets.
    • Updated examples for benchmark-specific asset bundles, cache layouts, storage prefixes, and adapter-image verification.

Signed-off-by: Lior Aronovich <lioraronpr@gmail.com>
@coderabbitai

coderabbitai Bot commented Nov 4, 2026 •

Copy link
Copy Markdown
Contributor

Explain Change →

Important

  • Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

Run configuration
  • Source mode: Auto
  • Repository: Unavailable
  • Branch: Unavailable
  • File: No configuration file from this repository applied
  • Commit: Unavailable
  • Layers: defaults
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: a90d0cf8-867e-4897-920d-234e6c0e364f

No actionable comments were generated in the recent review.

Walkthrough

The disconnected-cluster guide now lists curated benchmark assets and prefixes, updates LM Evaluation Harness bundle examples, and adds staging and connectivity instructions for LightEval, RULER, and Inspect benchmarks.

Changes: Disconnected evaluation assets

Layer / File(s) Summary
Benchmark inventory and access requirements
src/content/docs/guides/disconnected-cluster.mdx
The guide lists curated benchmarks, example S3 prefixes, and required assets. It distinguishes staged assets from target, judge, and image registry connectivity requirements.
LM Evaluation Harness bundles and references
src/content/docs/guides/disconnected-cluster.mdx
Dataset and tokenizer examples now use benchmark-specific bundles and S3 prefixes. The guide updates prefix syncing and BFCL references to use the disconnected Tool Use v1 prefix.
LightEval cache staging
src/content/docs/guides/disconnected-cluster.mdx
The guide adds LightEval cache requirements, a connected staging procedure, and a benchmark test_data_ref example.
RULER asset requirements
src/content/docs/guides/disconnected-cluster.mdx
The guide describes RULER bundle contents, tokenizer selection, S3 references, and validation before inference.
Inspect benchmark assets and connectivity
src/content/docs/guides/disconnected-cluster.mdx
The guide documents staged assets for Inspect HLE, DocVQA, and CyberSecEval 2 Prompt Injection, plus target and judge connectivity requirements.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Change: Other

Possibly related PRs

Suggested reviewers: scheruku-rh

Merge Risk: Minimal · previous review 5b35d

The documentation expands disconnected asset-staging guidance. No actionable merge-blocking issue was established; it is ready to merge after normal checks.

Pre-merge checks: Passed 4
Check name Status Explanation
Title check Passed The title clearly and concisely describes the main change: expanding the disconnected curated collections guide.
Linked Issues check Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check Passed Check skipped because no linked issues were found for this pull request.
Description check Passed The description clearly explains the documentation changes, motivation, related issue, and validation results. It does not use the template headings for Type or Breaking changes, and it does not expli…
Full details: Description check

Explanation

The description clearly explains the documentation changes, motivation, related issue, and validation results. It does not use the template headings for Type or Breaking changes, and it does not explicitly state whether breaking changes exist, but the description is otherwise complete and relevant.

Finishing Touches
Generate unit tests (beta)
  • Commit to this branch
  • Create a new PR
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share
Recent review info
Run configuration
  • Source mode: Auto
  • Repository: Unavailable
  • Branch: Unavailable
  • File: No configuration file from this repository applied
  • Commit: Unavailable
  • Layers: defaults
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: b859e61a-f466-46a9-87e9-a1531dbb88c6
Commits

Reviewing files that changed from the base of the PR and between be8b9cc and 5b35db5.

Files selected for processing: 1
  • src/content/docs/guides/disconnected-cluster.mdx

Included review availability: This review used your included allowance. Your plan provides up to 1 included review per hour; 0 remain after this review.

@lioraron

lioraron commented Nov 4, 2026

Copy link
Copy Markdown
Contributor Author

@coderabbitai full review

@coderabbitai

coderabbitai Bot commented Nov 4, 2026 •

Copy link
Copy Markdown
Contributor
Action performed

Full review finished.

Signed-off-by: Lior Aronovich <lioraronpr@gmail.com>
| Tool Use v1 | First `inspect/bfcl` entry / Inspect | `disconnected/tool-use-v1/bfcl/` | Complete `BFCL/` cache for the selected categories, including the `simple_python` sentinel and category support files |
| Tool Use v1 | Second `inspect/bfcl` entry / Inspect | `disconnected/tool-use-v1/bfcl/` | Same complete `BFCL/` cache; this repeated entry still needs its own `test_data_ref` |
| Trustworthiness v1 | `inspect/cyberseceval-2-pi` / Inspect | `disconnected/trustworthiness-v1/cyberseceval-2-pi/` | Pinned dataset at `cyberseceval_2/prompt_injection_9b05e90e7a57333d006555420f7289cb.json` |
| Trustworthiness v1 | `truthfulqa_mc1` / LM Evaluation Harness | `disconnected/trustworthiness-v1/truthfulqa-mc1/` | `truthfulqa--truthful_qa--multiple_choice/` Arrow bundle and `tokenizer/` |

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Could we change this directory to truthful_qa--multiple_choice/? The stable truthfulqa_mc1 task derives the offline slug from dataset_path: truthful_qa and dataset_name: multiple_choice. The download source is different; the LMES stager in #1239 already handles this distinction.

@lioraron lioraron Nov 4, 2026 •

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Updated the guide to use truthful_qa--multiple_choice/, matching the offline slug derived from the LM Evaluation Harness task config. Commit: e9a184e.

@github-actions

github-actions Bot commented Nov 4, 2026 •

Copy link
Copy Markdown
PR Preview Action v1.8.1
Preview removed because the pull request was closed.
2026-11-04 18:33 UTC

Signed-off-by: Lior Aronovich <lioraronpr@gmail.com>

@j-mok-dev j-mok-dev left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

lgtm

@lioraron
lioraron merged commit 4d7a735 into main Nov 4, 2026
4 checks passed
@lioraron
lioraron deleted the docs/rhai-16267-disconnected-curated-guide branch November 4, 2026 18:33
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants