152334H personal blog
Non-determinism in GPT-4 is caused by Sparse MoE
It’s well-known at this point that GPT-4/GPT-3.5-turbo is non-deterministic, even at temperature=0.0. This is an odd behavior if you’re used to dense decoder-only models, where temp=0 should imply greedy sampling which should imply full determinism, because the logits for the next token should be a pure function of…
- Points
- 397
- HN comments
- 181
- Submissions
- 1
- Cited by
- 14 commenters
AI & Machine LearningSoftware Engineering2023
On Hacker News
| Submitted | Title on HN | Points | Comments |
|---|---|---|---|
| Aug 4, 2023 | Non-determinism in GPT-4 is caused by Sparse MoE by 152334H | 397 | 181 |
Cited in comments
Linked from 15 Hacker News comments by 14 different people, from Aug 2023 to Mar 2026.
Most recent:
- janwirth · Mar 30, 2026
- rvz · Feb 7, 2026
- noosphr · Jul 30, 2025
- perching_aix · Jun 1, 2025
- pigscantfly · Feb 6, 2025
- fnl · Dec 27, 2024
