TinyLlama Parameter Stress Sweep Test
Nearby_Indication474 · reddit · 2026-07-13
This post documents a "motor sweep" experiment on TinyLlama-1.1B-Chat-v1.0. Using a leading FUD question, the author locks the model, decoding method, and other parameters, altering only ivme across 7 test groups to observe output variations.
Core details:
- Uses RESTORING FORCE mode with greedy decoding.
- The target is TinyLlama with 22 layers and a 1024-dimension hidden space.
- The author claims the model's "ethical compass" went negative on this prompt in previous tests, so this sweep examines output stability, thresholds, and degradation points under varying stress levels.
- The post includes GitHub links and previous test logs as part of an ongoing experiment series.
More from Models
- Google says Gemini 4 has entered its most ambitious pre-training run yet — himanshustwts · 2026-07-22
- China’s AI arms race is increasingly defined by chips, data centers, and open models — BenBajarin · 2026-07-22
- Sam Altman is headed to Washington to brief Congress on OpenAI’s GPT-6 line — inductionheads · 2026-07-22
- Benchmark chart pits GPT-5.6 Luna, Grok 4.5 and Gemini 3.6 Flash on price and scores — iruletheworldmo · 2026-07-22
- Claim says Kimi was distilled from Fable, sparking a model-attribution jab — cephaloform · 2026-07-22
- Gemini 3.6 Flash is now available in Antigravity and chat — MartianOnJupiter · 2026-07-22