@chip.gpt: This demonstrates how conversations with Large Language Models work. Every interaction is a silo and must include previous interactions. That is called the “context”. #ai #openai #gpt #chatgpt #claude #llm #mcp

Chip | Building DecRec
Chip | Building DecRec
Open In TikTok:
Region: US
Wednesday 30 July 2025 23:16:56 GMT
29892
1230
21
64

Music

Download

Comments

itsmemrfantastic
James™ :
so many people won't get this lol
2025-09-12 19:45:52
19
insightful_buzz
Insightful_Buzz :
ah the context window. good insight.
2025-07-31 02:01:22
181
depth_driven
Depth Driven :
This is exactly why hallucinations get worse with models in extended context scenarios, great example.
2025-08-11 17:54:25
177
iampeterprins
Peter Prins :
Most people don’t know this and if they did it would take some of the magic out of it for them
2025-07-31 04:02:00
96
rolifromhermes
Roli Bosch @ Hermes Labs :
this is why, when you start promoting about a different context or circumstance, you want to do so in a new, clean session. Otherwise, you'll have contextual spillover from everything it is carrying.
2025-08-22 22:37:57
45
realasdfasdfqwe
The Real asdfasdfqwe :
you forgot embedded hidden instructions from closedai
2025-08-22 17:13:11
31
ai.nurse
ai.nurse :
every message is basically a new chat
2025-07-31 03:09:15
27
msjcspztmxq
codykarinashared :
people say hi how are you to a llm? and don't just ask it the question straight up? wait do people do that in real life too? do they ask hey how are you before asking questions?
2025-09-20 17:26:23
1
llm.guy
LLM Guy :
Context windows 😅
2025-07-31 23:11:06
5
bsif47
Bsif :
This goes on anywhere from 4-8 thousand times
2025-09-18 04:33:39
3
alienbrill
Imbrilliant :
This is pretty much how regular human interaction works too
2025-09-23 05:01:41
1
theccsi
TheCCSI.org :
In the memory system I’m using, it’s just the last interaction. Then the query and a summary of the last interaction are used to search for relevant stored summaries of previous interactions that relate to our current interaction. “Just in time” context building seems to help.
2025-09-20 17:06:27
0
thelimester_
TheLimester_ :
Teams ping scared me
2025-09-04 05:49:19
1
drewstan1
NL_Vibe :
it’s just conversational language people overthink it …i suppose after thousands of hours immersed in Ai across multiple LLms sometimes at once it just flows with rhythm and they follow your vibe and nuance..
2025-08-08 22:36:19
1
supervyln
EJ :
This is a rag system
2025-09-17 09:26:28
0
serif_x
Ch :
Tbf humans need context too. Which makes me wonder… how would chat gpt interactions progress if we allow chat gpt to forget context like humans do all the time hah
2025-07-31 00:39:35
6
To see more videos from user @chip.gpt, please go to the Tikwm homepage.

Other Videos


About