Subliminal Learning Is More About Data Efficiency
AdtRaghunathan · x · 2026-07-09
This post discusses a conclusion regarding subliminal learning: the previous belief that it requires LoRA is inaccurate, and the core issue is more likely data efficiency. The author points out that with an appropriate learning rate and sufficient data, this phenomenon can occur even without using LoRA or FFT.
Related event: Subliminal Learning Hinges on Data Efficiency, Study Says(2 posts)→
More from Research
- GigaChat Audio targets long-form audio grounding with timestamps across 120-minute inputs — ai-sage · 2026-07-21
- Paper models Transformer components as stochastic geometry and tests five architectures — Zhihua Liang · 2026-07-21
- LTX 2.3 LoRA demo changes a video’s camera angle — CQDSN · 2026-07-21
- OpenForecaster uses daily news to improve language-model forecasting — Cohere_Labs · 2026-07-21
- Baseten study finds new facts in LLM weights are fragile unless trained from many restatements — alex_verem · 2026-07-21
- uv-scripts/ocr returns to the top of Hugging Face datasets with a JSON model picker — vanstriendaniel · 2026-07-21