Post
5402
๐งฌ Darwin Family: Zero Gradient Steps, GPQA Diamond 88.89%
How far can we push LLM reasoning *without* training?
Our team at VIDRAFT submitted this paper to Daily Papers yesterday, and it's
currently #3. Huge thanks to everyone who upvoted โ sharing the core ideas below.
๐ Paper: Darwin Family: MRI-Trust-Weighted Evolutionary Merging for Training-Free Scaling of Language-Model Reasoning (2605.14386)
๐ arXiv: https://arxiv.org/abs/2605.14386
๐ Model: FINAL-Bench/Darwin-28B-REASON
๐ Model: FINAL-Bench/Darwin-28B-Opus
---
TL;DR
Darwin Family is a training-free evolutionary merging framework.
By recombining the weight spaces of existing LLM checkpoints โ with zero
gradient-based training โ it reaches frontier-level reasoning.
- ๐ Darwin-28B-Opus: GPQA Diamond 88.89%
- ๐ธ Zero gradient steps โ not a single B200 or H200 hour needed
- ๐งฌ Consistent gains across 4B โ 35B scale
- ๐ Cross-architecture breeding between Transformer and Mamba families
- ๐ Stable recursive multi-generation evolution
#Three Core Mechanisms
โ 14-dim Adaptive Merge Genome โ fine-grained recombination at both
component level (Attention / FFN / MLP / LayerNorm / Embedding) and block
level, expanding the prior evolutionary-merge search space.
โก MRI-Trust Fusion โ we diagnose each layer's reasoning contribution
via an **MRI (Model Reasoning Importance)** signal and fuse it with
evolutionary search through a **learnable trust parameter**. Trust the
diagnostic too much and search collapses; ignore it and search becomes
inefficient โ Darwin learns the balance from data.
โข Architecture Mapper โ weight-space breeding across heterogeneous
families. Attention ร SSM crossover actually works.
Why It Matters
> Diagnose latent capabilities already encoded in open checkpoints,
> and recombine them โ no gradients required.
Replies and critiques welcome ๐
How far can we push LLM reasoning *without* training?
Our team at VIDRAFT submitted this paper to Daily Papers yesterday, and it's
currently #3. Huge thanks to everyone who upvoted โ sharing the core ideas below.
๐ Paper: Darwin Family: MRI-Trust-Weighted Evolutionary Merging for Training-Free Scaling of Language-Model Reasoning (2605.14386)
๐ arXiv: https://arxiv.org/abs/2605.14386
๐ Model: FINAL-Bench/Darwin-28B-REASON
๐ Model: FINAL-Bench/Darwin-28B-Opus
---
TL;DR
Darwin Family is a training-free evolutionary merging framework.
By recombining the weight spaces of existing LLM checkpoints โ with zero
gradient-based training โ it reaches frontier-level reasoning.
- ๐ Darwin-28B-Opus: GPQA Diamond 88.89%
- ๐ธ Zero gradient steps โ not a single B200 or H200 hour needed
- ๐งฌ Consistent gains across 4B โ 35B scale
- ๐ Cross-architecture breeding between Transformer and Mamba families
- ๐ Stable recursive multi-generation evolution
#Three Core Mechanisms
โ 14-dim Adaptive Merge Genome โ fine-grained recombination at both
component level (Attention / FFN / MLP / LayerNorm / Embedding) and block
level, expanding the prior evolutionary-merge search space.
โก MRI-Trust Fusion โ we diagnose each layer's reasoning contribution
via an **MRI (Model Reasoning Importance)** signal and fuse it with
evolutionary search through a **learnable trust parameter**. Trust the
diagnostic too much and search collapses; ignore it and search becomes
inefficient โ Darwin learns the balance from data.
โข Architecture Mapper โ weight-space breeding across heterogeneous
families. Attention ร SSM crossover actually works.
Why It Matters
> Diagnose latent capabilities already encoded in open checkpoints,
> and recombine them โ no gradients required.
Replies and critiques welcome ๐
- 3 replies
