How I Built an AI Scholarship Finder with Olostep
kingabzpro
• • 2NEW Articles from Team or Enterprise organizations will get promoted to the main section.
The Hold, the Seam, the Turn
ChromaticCoherence
• • 1Catching Confident Hallucinations from the Inside: A Linear Probe on a Model's Own Activations, and Where It Stops Working
assafpet
• Hidden-State Hallucination Probes Beat Confidence Baselines — And the Gap Grows with Scale
sarimahsan101
• CiberIA Expected Cognitive Profile Fast: Mythos - Fable
gcjordi
• Claude Fable 5 — Technical Harness Report
Svngoku
• • 3Pulpie: Pareto-Optimal Models for Cleaning the Web
feyninc
• • 4AstroBERT Small: Domain-specialized small models
NeuML
• • 5CiberIA Expected Cognitive Profile Fast v1.0: Claude Sonnet 5
gcjordi
• BaseRT: Best-in-Class LLM Inference on Apple Silicon via Native Metal
basecompute
• • 1Left vs. Right Alignment, A case study from porting LeRobot PI05's subtask prediction from JAX to PyTorch
jorgemunozl
• • 1Does Your LLM Know *When It's About to Be Wrong*?
ginigen-ai
• • 17What if you cached the model's hidden states instead of running it again?
luizspies
• Heretic Grimoire
VINAY-UMRETHE
• • 5Building an agent that researches competitors and ships Meta ad creatives
mesmertech
• • 1Teaching Machines to Read Silicon: An Open, Generated Dataset of SoC RTL
hasankursun
• • 180TB+ of astronomy for the HDD-poor: crossmatch the Multimodal Universe from your laptop
hugging-science
• • 17Gemma-4 31B + vLLM on RTX 6000 PRO : A Real-Load Benchmark
hexgridcloud
• • 3🔧 L'architecture est un seuil, pas un levier — ce que j'ai appris en optimisant un LLM français de 15M de paramètres 🇫🇷
RDTvlokip
• • 1🔧 Architecture is a threshold, not a lever — what I learned optimizing a 15M French LLM 🇫🇷
RDTvlokip
• • 1