GenAI Stack Exchange

7 votes

2 answers

447 views

Why should one avoid adding a system prompt with DeepSeek-R1?

Rob

126

answered 1 hour ago

0 votes

1 answer

2 views

What is the pattern for multiple globs in applyTo for custom instruction files in github copilot

bradgonesurfing

101

answered 9 hours ago

9 votes

3 answers

4k views

What are the limitations that prevent LLM's from continuing their response

Franck Dernoncourt

4,797

answered 2 days ago

1 vote

2 answers

49 views

How was the claim "Gemini 2.0 Flash achieves 24x higher intelligence per dollar than anyone in the market" determined?

JSM-101

21

answered 2 days ago

3 votes

2 answers

229 views

Separating information in prompt

Kathryn S

41

answered May 6 at 7:38

1 vote

5 answers

950 views

Is it feasible to restrict a large language model (LLM) to respond exclusively within a user-specific context?

Future AGI

1

answered May 6 at 7:24

5 votes

1 answer

254 views

How to fine-tune an LLM on a very small document

Kathryn S

41

answered May 6 at 7:02

-1 votes

1 answer

23 views

Implementing copilot-like feature: What technologies/APIs to use?

Kathryn S

41

answered May 6 at 6:24

0 votes

0 answers

6 views

Curious about tech-stack choices when creating chat-bots

Kathryn S

41

asked May 6 at 4:05

0 votes

0 answers

7 views

ChatGPT App on Android Stops too often

NealWalters

101

asked May 5 at 20:59

1 vote

0 answers

31 views

Is there any point in using GPT o1 now that o3 is available and cheaper?

Franck Dernoncourt

4,797

modified May 3 at 0:06

1 vote

1 answer

66 views

Can a model still imitate modern characters if finetuned exclusively using public domain characters as reference system prompts?

Franck Dernoncourt

4,797

modified May 2 at 0:47

0 votes

0 answers

11 views

How to shorten "Please <action>" prompts, while maintaining neutral tone?

Jason R. Mick

101

modified May 1 at 21:29

2 votes

1 answer

50 views

Why does ChatGPT often answer a question I did not ask?

Franck Dernoncourt

4,797

answered May 1 at 6:07

Stack Exchange Network

Explore our questions

Why should one avoid adding a system prompt with DeepSeek-R1?

What is the pattern for multiple globs in applyTo for custom instruction files in github copilot

What are the limitations that prevent LLM's from continuing their response

How was the claim "Gemini 2.0 Flash achieves 24x higher intelligence per dollar than anyone in the market" determined?

Separating information in prompt

Is it feasible to restrict a large language model (LLM) to respond exclusively within a user-specific context?

How to fine-tune an LLM on a very small document

Implementing copilot-like feature: What technologies/APIs to use?

Curious about tech-stack choices when creating chat-bots

ChatGPT App on Android Stops too often

Is there any point in using GPT o1 now that o3 is available and cheaper?

Can a model still imitate modern characters if finetuned exclusively using public domain characters as reference system prompts?

How to shorten "Please <action>" prompts, while maintaining neutral tone?

Why does ChatGPT often answer a question I did not ask?

Hot Network Questions