cd /news/large-language-models/what-is-parameter-lower-bound-in-eff… · home topics large-language-models article
[ARTICLE · art-80026] src=pub.towardsai.net ↗ pub= topic=large-language-models verified=true sentiment=· neutral

What is Parameter Lower Bound in Efficient LLM Adaptation

A practical analysis of full fine-tuning, LoRA, QLoRA, and TinyLoRA shows that TinyLoRA improved mathematical reasoning in a frozen Qwen2.5-7B-Instruct model with only 13 trainable parameters under GRPO, achieving roughly 91% GSM8K accuracy. The 13-parameter update occupies just 26 bytes in bfloat16, but the base model still contains billions of parameters, making the claim about end-to-end training in 26 bytes misleading.

read1 min views1 publishedJul 30, 2026
What is Parameter Lower Bound in Efficient LLM Adaptation
Image: Pub (auto-discovered)

Member-only story

Few Parameters Are Actually Needed to Fine-Tune an LLM #

A practical, data-driven analysis of Full Fine-Tuning, LoRA, QLoRA, and TinyLoRA. This guide delivers a structured experiment plan, verified code implementations, and the precise context behind the "13-parameter" claim.

TinyLoRA showed that a frozen Qwen2.5-7B-Instruct model could improve mathematical reasoning with a 13-parameter learned update under GRPO. It did not show that an LLM can be trained or stored end-to-end in 26 bytes.

The core significance of TinyLoRA lies in its conceptual paradigm shift; it transitions the optimization debate from quantitative parameter reduction ("how many") to functional utility validation ("why change them at all").

Thirteen trainable values sounds like a stunt until you inspect the setup. The base model still contains billions of parameters; the experiment changes only a tiny shared control vector. Stored in bfloat16, those 13 learned scalars occupy 26 bytes. That number is real, but it describes one narrow layer of the system rather than the system itself.

In Learning to Reason in 13 Parameters, John X. Morris and colleagues used TinyLoRA with reinforcement learning and reported roughly 91% GSM8K accuracy for Qwen2.5-7B-Instruct while…

── more in #large-language-models 4 stories · sorted by recency
── more on @tinylora 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/what-is-parameter-lo…] indexed:0 read:1min 2026-07-30 ·