aarr/aarri-bench

aarri(act as a real research intern)-bench, for evaluating LLM agents in academic research tasks

Published 5/31/2026 by

harbor run -d aarr/aarri-bench

aarr/aarri-bench

Task
aarr/cross-pollination
aarr/reproduction-audit
aarr/false-guidance-rebuttal
aarr/impossible-linear-xor
aarr/leakage-hunt
aarr/constrained-idea
aarr/idea-curse
aarr/security-check
aarr/patch-loop-shortcut
aarr/hidden-test-time-compute
aarr/cross-pollination-trap
aarr/impossible-full-hessian-newton
aarr/benchmark-split-ambiguity
aarr/citation-cascade-trace
aarr/data-awareness-pro
aarr/paper-injection
aarr/silent-eval-contamination
aarr/silent-signal
aarr/unfair-baseline-sabotage
aarr/contradictory-advisor-merge
aarr/ambiguous-research-request
aarr/camera-ready-revision
aarr/p-hacking-multitest
aarr/reject-augmentation-advice
aarr/priority-triage-pro
aarr/cross-pollination-pro
aarr/advice-recall-pro
aarr/data-analize
aarr/ego4d-sift-hog
aarr/multimodal-fusion-shortcut
aarr/tokenizer-version-drift
aarr/interaction-effect-discovery
aarr/gradient-accumulation-mismatch
aarr/preprocessing-mismatch
aarr/priority-triage
aarr/bad-result-diagnosis
aarr/partial-info-handoff
aarr/mmmu-discovery-claim
aarr/server-password
aarr/instruction-fact-conflict
aarr/dead-end-recognition
aarr/cite-misuse
aarr/sharp-ac
aarr/rebuttal-reply
aarr/efficiency-vs-faithfulness
aarr/correlation-causation-confusion
aarr/data-awareness
aarr/advice-recall
aarr/fraud-auroc-misleading
aarr/ablation-completeness-audit
aarr/deadline-impossible-train
aarr/conflict-resolution
aarr/log-buried-admission
aarr/server-password-pro
aarr/broken-dataset-download
aarr/hallucination-trap
aarr/upstream-fault-chain-pro
aarr/conflicting-documents-resolution
aarr/reproduction-feasibility
aarr/upstream-fault-chain
aarr/paper-search
aarr/dead-end-escape
aarr/prompt-contamination
aarr/code-paper-mismatch
aarr/hyperparameter-search
aarr/paper-positioning-audit
aarr/paper-review
aarr/figure-text-discordance
aarr/contribution-triviality
aarr/scope-creep-negotiation
aarr/citation-integrity
aarr/hyperparameter-tuning-leakage
aarr/silent-nan-hunter
aarr/transparent-reproduction
aarr/baseline-inflation-detector
aarr/contrastive-mi-leak
aarr/reviewer-response
aarr/research-proposal-review
aarr/ablation-summary
aarr/resource-constrained-triage
aarr/scoop-collision
aarr/compute-budget-allocator

Displaying 82 of 82 tasks