research·linter

Lint report · stored run

CUDA-Harness: Harnessing Agentic CUDA Kernel Generation and Optimization from Natural Language

Qi Fan, An Zou, Yehan Ma open the source paper via the report id, not the extraction source: pdf IR v1.0 generated 2026-09-14 04:12:39 UTC
Lint another paper Raw report JSON
0errors
0warnings
0info
23claims read
283numbers extracted
30references
How well we read this paper 1 observation

These are notes about our reading of the paper, not about the paper. They affect how much weight the findings below can carry.

Note 1
  • compile-warning references: the reference list carries no entry numbers; citation keys are enumeration order [1]..[N], not labels printed in the paper, so a key may not match the [N] used in the body
Some checks could not run
  • RL002 reported nothing, and it could not reach OpenAlex — so its silence here is not evidence that this paper is clean.
  • RL003 reported nothing, and it could not reach OpenAlex — so its silence here is not evidence that this paper is clean.

Everything else on this page was computed from the paper itself and is unaffected. This is a limit of this run, not a finding about the paper.

Compiler extraction log (1) 
references: the reference list carries no entry numbers; citation keys are enumeration order [1]..[N], not labels printed in the paper, so a key may not match the [N] used in the body

No findings on this paper.

That is a result, not a failure. The compiler read the paper without errors, and no rule found something it could prove with evidence — this product would rather say nothing than say something vague.

Read the extraction notes above before taking this as a clean bill of health: they say which checks were working with incomplete information.

Rule notes (36) 
RL001 funnel arithmetic: sentences_with_from=2 endpoint_pairs=0 claims_compared=0 emitted=0 [dropped: not-a-measurement=0]
RL001 funnel delta-column: structured_tables=3 with_exactly_one_delta_column=0 rows_checked=0 emitted=0
RL001 funnel cross-location: quant=283 prose=122 prose_with_metric=4 prose_with_dataset=0 prose_with_metric_and_dataset=0 distinct_buckets=0 emitted=0 [dropped: not-a-measurement=1 not-in-quote=20]
RL001 funnel identifier-drift: settings_checked=2 sections_stating_one_value=0 emitted=0
RL002 citation entailment: attempted 11 of 11 ranked cited claims (36 in-text markers from 30 bibliography entries); 0 cited works could not be identified, 2 had no retrievable text, 0 model quotes failed verbatim verification, 0 were topically unrelated to the retrieved text.
RL002 candidate ranking (why the top 11 were chosen): [27] [introduction, absolutist wording, title/sentence overlap near zero]; [11] [introduction, names an artefact (ConCuR)]; [8] [introduction, names an artefact (ReGraphT)]; [16] [introduction, title/sentence overlap near zero]; [5] [introduction, title/sentence overlap near zero]; [1] [experiment, names an artefact (Besides Seed2.0 Lite), absolutist wording, title/sentence overlap near zero, no DOI/arXiv id]; [25] [experiment, names an artefact (GLM-5.1), absolutist wording, title/sentence overlap near zero]; [20] [experiment, names an artefact (KernelSkill), absolutist wording, title/sentence overlap near zero]
RL002 label distribution over 9 classified citations: PARTIALLY_SUPPORTED=1, SUPPORTED=8
RL002 base-rate self-check: not tripped (0/9 = 0% NOT_SUPPORTED, threshold 40%).
RL002 attribution check (case B): 5 candidates name an artefact, 5 of them made the top-N and were checked against their cited source
RL002: 1 citations skipped because the sentence reports this paper's own results (the citation there is a comparison pointer, not evidence for the sentence)
RL002: 3 enumerated-citation sentences skipped (they cite >=3 works, so the claim is about the set and cannot be refuted by any single cited work)
RL002: 1 attributions dropped as family variants (the cited work is an adjacent version in the same family, e.g. 'DeepSeek-Prover' against DeepSeek-Prover-V2) -- not cleared, just not worth an author's time
RL003: 3 queries -> 93 unique works (arXiv search 44, category listings 0, scholarly graph 49 [crossref 49], 0 graph works matched to an arXiv id); source errors during retrieval: arXiv 0, graph 0
RL003: parsed the reference list of 9/12 comparable papers (3 unreadable) — that is the PeerUsage denominator
RL003: enriched 1 of 3 candidates with their own abstract (2 unavailable, 0 over budget); 0 dropped as published after this paper
RL003: scholarly graph for this run = crossref
RL003 funnel (300 candidates): already_used_by_the_paper -9, cited_only_from_related_work -79, is_the_paper_itself -1, peer_usage_below_quorum -86, survey_or_artifact -43, title_names_no_method -79 || kept: citation_context_unknown +45 || survived-cheap-gates 3 -> with-abstract 1 -> comparable 3 -> above-threshold-0.65 0
RL003 candidate: 'Kevin' score 0.636 (T 0.50 D 0.50 M 1.00 P 0.56 R 1.00; 5/9 peers cite it) -> not reported
RL003 candidate: 'Dr. Kernel' score 0.442 (T 0.00 D 0.50 M 1.00 P 0.33 R 1.00; 3/9 peers cite it; title-only, no abstract available) -> not reported
RL003 candidate: 'Triton' score 0.399 (T 0.00 D 0.50 M 1.00 P 0.33 R 0.14; 3/9 peers cite it; title-only, no abstract available) -> not reported
RL003: 3 comparable candidates all scored below 0.65 — nothing to report for this paper
RL004 funnel: components=12 contribution_backed=6 candidates=3 ablations_found=5 matched_by_ablation=3 unmatched=0 experiments=7
RL004 component M1 'Intermediate-Structured Generation': covered (candidate)
RL004 component M2 'Kernel Implementation Manifest': covered (filtered earlier)
RL004 component M3 'Self-Contained CUDA Scaffold': covered (filtered earlier)
RL004 component M4 'Synthesis-Based Verification': covered (candidate)
RL004 component M5 'Decoupled Test Data Synthesis': covered (filtered earlier)
RL004 component M6 'Progressive Multi-Stage Validation': excluded:stage-of-a-larger-claimed-component
RL004 component M7 'Feedback-Adaptive Evolution': covered (candidate)
RL004 component M8 'Correctness-First Refinement': covered (filtered earlier)
RL004 component M9 'Reward-Driven Optimization': covered (filtered earlier)
RL004 component M10 'optimization insight summarization': excluded:not-a-claimed-contribution
RL004 component M11 'Reinforcement Learning with Verifiable Reward (RLVR)': excluded:not-a-claimed-contribution
RL004 component M12 'reasoning-action (ReAct) paradigm': excluded:not-a-claimed-contribution
RL004 funnel: ranked=0 emitted=0 suppressed_by_cap=0 dropped_unverified_quote=0
RL005: repo https://github.com/anomalyco/opencode unusable (stopped early after 214 files / 4008190 bytes)
Reading the report as data

This page is a rendering of the report the pipeline wrote. The JSON below is the stored artefact, unedited — the same document research-lint --format json produces. 

loading…