How Will the LLM Hallucination Problem Be Solved?

1kṀ965

2029

11%

Other

Vector Embeddings (as with Pinecone https://www.pinecone.io/)

1.2%

Filtering (as with Deepmind AlphaCode https://www.deepmind.com/blog/competitive-programming-with-alphacode)

0.1%

Ensemble Combined with Fine Tuning

0.5%

Joint Embedding Predictive Architecture (https://arxiv.org/pdf/2301.08243.pdf)

0.3%

Feed Forward Algorithms (https://www.cs.toronto.edu/~hinton/FFA13.pdf)

10%

Bigger model trained on more data + RL

1.6%

Vigger models + prompt engineering

47%

It won't be

20%

Giving all LLMs access to the internet and databases of scientific papers

By the year 2028, how will the Hallucination Problem have been solved for the vast majority of applications out there?

— LLM & AI Capabilities—

Get

1,000

to start trading!

People are also trading

Will LLM hallucinations be a fixed problem by the end of 2025?

6% chance

Will LLM hallucinations be a fixed problem by the end of 2028?

45% chance

LLM Hallucination: Will an LLM score >90% on SimpleQA before 2026?

60% chance

Will hallucinations (made up facts) created by LLMs go below 1% on specific corpora before 2025?

41% chance

Will scaling current methods be enough to eliminate LLM hallucination?

15% chance

Do LLMs experience qualia?

29% chance

How will the data shortage for LLM gets solved

Will an LLM be able to solve Raven's Progressive Matrices from an image in 2025?

65% chance

Will LLMs be banned at the 2026 MIT Mystery Hunt?

18% chance

Will LLMs mostly overcome the Reversal Curse by the end of 2025?

Sort by:

I'm guessing models will eventually be taught to express uncertainty in their answers. When asked "who is [person that does not exist]" many of them already correctly say they don't know who they are. I see no theoretical reason this capability couldn't be broadened

@LukeFrymire this would imply that either the options in my market are overvalued or your market is overvalued.

@PatrickDelaney I think the resolution criteria is fairly different. Mine requires that a scale-based solution is possible, yours requires it to be the primary method in production.