Academic Publishing.
Zero Bullshit. Non-Profit.
Anyone can publish anything for a one-time flat fee of 5 CHF / USD / EUR. Your pre-print is accessible to anyone and everyone for free. Forever.
Upload your paper and pay the 5 flat fee. Accessible to everyone for free, forever. Reaching 5 citations automatically promotes your paper to "Accepted - Published".
For either the first 6 months or until reaching 10 additional citations (15 total), your paper can be voided if academic fraud (plagiarism, unchecked AI, fake data) is verified.
After Step 2 conditions pass, paper cannot be voided. If academic dishonesty is reported, a public warning badge is assigned and automatically propagates to any citing papers.
Cost-Aware Human-LLM Collaboration for post-OCR Corrections in Swiss Historical Newspapers
OCR transcription errors in historical archives often hinder digital search and retrieval. While Large Language Models (LLMs) can correct many of these errors, applying them indiscriminately is costly and may negatively affect already-clean text. We propose a three-tier collaboration framework that routes each text segment to one of: (1) No Correction, (2) LLM Correction, or (3) Human Correction. We introduce a regression-guided routing approach that prioritizes segments by predicted CER improvement, paired with a safeguard layer that detects harmful LLM corrections and routes uncertain segments to human review. With only <5% of the corpus reviewed by human experts, our safeguard achieves a 14% relative reduction over the All-LLM baseline, and substantially outperforms standard confidence-based approaches. By dynamically routing degraded segments to humans and fixable errors to the LLM, the collaborative framework outperforms either corrector in isolation.