Humanlaya @Humanlayadata

Built by model builders, for model builders. Humanlaya engineers expert-level data and verifiable rewards for frontier AI labs. humanlaya.com Joined March 2026

Tweets

16
Followers

148
Following

24
Likes

2

Humanlaya @Humanlayadata

2 months ago

GPT-5.6 is a growth cheat code! 🚀 It instantly turns complex metrics into actionable data dashboards. Deep analysis now takes seconds. 📊

0 0 1 360 1

View Details

Humanlaya @Humanlayadata

6 months ago

$OneMillion-Bench is our step toward that. See our work at：📄 <paper> arxiv.org/abs/2603.07980 | 💾 <data/code> github.com/humanlaya/OneM… | 📝 <blog> lab.humanlaya.com | 🤗 <hf>huggingface.co/datasets/Human…

0 2 4 275 0

View Details

Humanlaya @Humanlayadata

6 months ago

Why it matters + resources 🔗 If we want agents in finance/law/medicine/engineering, we need evaluation that matches real workflows + real cost of mistakes.

1 0 3 289 0

View Details

Dollars Matter, Scores Flatter. 💰 Can LLM agents really do high-valued experts work end-to-end — not just “solve a question”? Introducing $OneMillion-Bench benchmarks real expert work, priced in dollars 💵

3 3 14 2K 2

View Details

Humanlaya @Humanlayadata

6 months ago

What moves the needle 🔧 Tools and evaluation design matter a lot. • Official scaffolds > OpenRouter > no search (consistent gains) • Rankings are stable across different judge models (robustness check) • We also stress-test temporal sensitivity and test-time scaling behavior.

0 0 3 145 0

View Details

Humanlaya @Humanlayadata

6 months ago

How Does We Grade? ✅ ❌ We only count value when the output is deliverable, not “partially correct.” • 15–35 expert rubrics per task + negative rubrics for hallucination/unsafe/non-compliance • Asymmetric weights (-20 to +10) to reduce reward hacking

1 0 3 204 0

View Details

Humanlaya @Humanlayadata

6 months ago

From high-economic-value tasks to verifiable rewards, we help frontier teams push model capability beyond benchmark optimization and toward real-world cognitive performance.

0 0 2 101 0

View Details

Humanlaya @Humanlayadata

6 months ago

We design expert-level tasks, build rigorous verification systems, and turn high-density human knowledge into scalable training signals for advanced models.