In the strongest example of "weight space programming", as the technique is called, Astra’s intervention raised Qwen3.5-0.8B’s MATH-500 score from 7.9% to 48.1% while touching approximately 0.016% of ...

Article by Shenchao TechFlowNo training required, no data needed, no gradient descent.Researchers at MIT fed the weights of ...

As AI remakes the field of math, some students say they feel like giving up.

Vonovia trimmed 2026-2028 targets, citing higher financing costs, softer disposals and slower rent growth; Q3 report due ...

A new MIT study finds that many young mathematicians are reconsidering their future in the field as AI increasingly solves ...

What happens when one AI tries to fix another? MIT researchers used OpenAI's GPT-6 Astra to modify Alibaba's Qwen AI model ...

Young mathematicians are getting increasingly worried about their future as AI gets better at solving complex maths problems.

MIT researchers just gave Alibaba's Qwen AI model a serious math upgrade using GPT-6 Astra. By tweaking only 0.016% of its neural network, Astra helped Qwen ditch repetitive thinking and solve ...

Those who trust data and probability over experience and intuition will win.”Having studied statistics and probability theory during my student days, I have long been drawn to the “fascination of ...

MIT CSAIL researchers say GPT-6 Astra edited 0.016% of Qwen3.5-0.8B weights to fix reasoning loops, lifting MATH-500 from 7.9% to 48.1% without training.

Qwen3.5-0.8B, a small open-weight model from Alibaba, also saw GSM8K accuracy rise from 13.8% to 67.1% and IFEval from 21.0% to 62.3%. The researchers call the result early "sparks" of a capability, ...

MIT has unveiled an innovative, groundbreaking paradigm dubbed "Weight Space Programming", representing a transformative new ...