AI Paper on Sliding-Window Attention Draws Criticism
Replies target an arXiv paper on sliding-window attention with sinks.
A reply posted by @teortaxesTex calls out an arXiv paper that claims sliding-window attention with sinks outperforms prior methods. The post notes that the paper's title and the replies it received attracted more notice than most similar releases. Typical papers of this kind receive little sustained attention and fade quickly even when they gather thousands of likes. The author frames the current case as an exception worth examining because of how the title and ensuing comments unfolded. No independent confirmation of the paper's technical claims appears in the supplied lines.
Some of the most memorable papers have interesting/funny titles while still being honest about their work In fact, if one has to resort to clickbaity titles that are unfaithful to the work done, it is likely because the work itself is not interesting or engaging enough
@giffmana @JFPuget @RheaSukthanker @CameronPashmina That would have cut our visibility by 90%. The fact is, you saw the title, you engaged with it. It worked its purpose. All you bums can do is criticize the presentation because the content is impeccable.
Reactions
2 postsT( 
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)@teortaxesTexinsane response that made everything 10x worse. what are we even doing here > Scientific Paper Consultant V4-Flash is better at your job than you are. Retract and apologize to mitigate the damage.
- likes: 153
- replies: 11
- bookmarks: 32
- reposts: 1
T( Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)@teortaxesTexlike nobody cares much if you publish a slop paper, more where that came from, whatever, will be forgotten in a week, 2k likes or no THIS, though…
- likes: 24
- replies: 0
- bookmarks: 0
- reposts: 0
