PulseAugur
EN
LIVE 00:56:26

MAMMOTH-1 leads LLM decision model tests; Qwen3.6, Gemma 4 excel at prompt injection

By PulseAugur Editorial · [1 sources] ·

A comparison of four decision models and five small large language models (LLMs) across 6,662 decisions revealed that MAMMOTH-1 performed best on average. However, Qwen3.6 and Gemma 4 demonstrated strong performance, both achieving 92.9% accuracy on prompt injection tasks. The researcher has made the testing harness and raw data publicly available, inviting suggestions for future tests. AI

IMPACT Provides insights into the comparative performance of smaller LLMs on decision-making and security vulnerabilities.

RANK_REASON Comparison of LLM performance on decision tasks and prompt injection. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Mastodon — mastodon.social →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

MAMMOTH-1 leads LLM decision model tests; Qwen3.6, Gemma 4 excel at prompt injection

How we ranked this

Signal score
10 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Comparison of LLM performance on decision tasks and prompt injection. [lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
model release, other
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. Mastodon — mastodon.social TIER_1 English(EN) · pournasserian ·

    I ran four decision models and five small LLMs on 6,662 decisions across six public datasets. Mammoth-1 led the average, but Qwen3.6 and Gemma 4 each reached 92

    I ran four decision models and five small LLMs on 6,662 decisions across six public datasets. Mammoth-1 led the average, but Qwen3.6 and Gemma 4 each reached 92.9% on prompt injection. I published the harness and raw calls. Which task would you test next? # AI # LLM https:// pour…

    LINKS pournasserian.com/…/decision-models-vs-sm…

RELATED ENTITIES

RELATED TOPICS