- 13 Posts
- 30 Comments
5·2 months agoBoy, I can’t wait for “Depthfind” 🥸 🐬
I also wonder what this would mean for LLMs that use Deepseek’s techniques…
This raises another question, LLM of Theseus
If I take Deepseek and alter it’s source code, will it still be considered “Deepseek”, at what point will it stop being Deepseek? If it’s not considered Deepseek, will I be able to run a service with it?
Notable OpenAI’s O1 cot is hidden with only a summary available sometimes
From Deepseek:
Chain of Thought (CoT) in LLMs refers to a prompting technique that guides large language models to articulate intermediate reasoning steps when solving a problem, mimicking human-like logical progression. Here’s a concise breakdown:
-
Purpose: Enhances performance on complex tasks (e.g., math, logic, commonsense reasoning) by breaking problems into sequential steps, reducing errors from direct, unstructured answers.
-
Mechanism:
- Prompting Strategy: Users provide examples (few-shot) or explicit instructions (zero-shot, e.g., “Let’s think step by step”) to encourage step-by-step explanations.
- Output Structure: The model generates tokens sequentially, simulating a reasoned pathway to the answer, even though internal processing remains parallel (via transformer architecture).
-
Benefits:
- Accuracy: Improves results on multi-step tasks by isolating and addressing each component.
- Transparency: Makes the model’s “thinking” visible, aiding debugging and trust.
-
Variants:
- Few-Shot CoT: Examples with detailed reasoning are included in the prompt.
- Zero-Shot CoT: Direct instructions trigger step-by-step output without examples.
- Self-Consistency: Aggregates answers from multiple CoT paths to select the most consistent one.
-
Effectiveness: Particularly impactful for tasks requiring structured reasoning, while less critical for simple queries. Research shows marked accuracy gains in benchmarks like math word problems.
In essence, CoT leverages the model’s generative capability to externalize reasoning, bridging the gap between opaque model decisions and interpretable human problem-solving.
Example:
-
Lol, what’s ClosedAI going to do? Hide ChatGPT’s answers behind a summary (like with the CoT) 🤣
The Chinese New Year robot situation is crazy
Footage of the crash: https://twstalker.com/WarMonitors/status/1884427724623626528
What’s ppls thoughts on Deepseek-R1, while I’ve heard chatter that it is a challenger to OpenAI O1 I’m not super knowledgeable on the technicalities of LLMs, but I’ve tried it, and it is fascinating to see its thought chain.
Holy fuck, the more I read in to the bot campaign, the more infuriated and horrified I get
“As it relates to Covid-19 disinformation, China [in 2020] initiated a disinformation campaign to falsely blame the United States for the spread of Covid-19,” Lawrence noted. “In line with the US National Defense Strategy, the DoD continues to build integrated deterrence against critical challenges to US national security, including deterring the PRC’s spread of disinformation under the scrutiny of the Department’s coordination and deconfliction process,” she said.
They are literally using “Whataboutism”, except spreading vaccine denial actually gets people KILLED.
According to the document, the Pentagon also conceded it had “made some missteps in our COVID related messaging” but assured the Philippines that the military “has vastly improved oversight and accountability of information operations” since 2022.
Oops-poopsies, we promise we won’t cause suffering and death the next time we deploy our bot army because we’re “Transparent” and “Accountable” (especially after we allotted 1.6 Billion dollars to badmouth China) Tee-hee-hee
Who gives a flying fuck about “spooky wumaos” when the US bot army actually got people KILLED
Oh yeah, that Vaccine misinformation one was the recent one that was publicly acknowledged, so there could be a lot of other bot/trolls armies we don’t know about :(
37·2 months agoHonestly, watching libs spew this bs cope mantra is so goofy and funny
This him?
deleted by creator
Honestly, I just really hope that this makes “woke” Sinophobes out themselves faster and show how little they know of (and care about) China outside of “muh Tinyman square” and other talking points
(I know she’s low hanging fruit, but you get the picture)
Interestingly enough, I think I saw some Indian users there.
Just like when the CPC temporarily teamed with the KMT to stop the Japanese invasion.
Except, correct me if I’m wrong, I don’t think Hamas has been actively anti-Communist unlike the KMT right?
This one is bothering me in particular. I don’t think any demands would be particularly revolutionary or anti-imperialist, it would just be rudimentary like demanding the U.S to hold the officials responsible for the campaign accountable and punish them (Since there doesn’t seem to be any indication that the officials responsible were even disciplined or demoted) something that I can see bog-standard liberals doing.
I’ve also done some other research about anti-US protests (at least in the US embassy) in the Philippines and found:
2014? - 2016 Protests against US troops in the Philippines
2022 Protests against Biden supporting the newly elected president Marcos due to the elections being corrupt and rigged
2023-2024 Pro-Palestine Protests
2025 - Anti-Trump Protests.
(I might have missed some others, I’ll wait for Philippine users to correct me)
What is most interesting is that there doesn’t seem to be any more recent anti-US military protests in the Philippines, even though the US has further expanded it’s presence there recently.
I think this is indicative of how US uses propaganda. The US doesn’t care how the Philippines views it, it only seems to care that it can use the Philippines as a stepping stone for its imperialism and military presence.
It seems to focus more on slandering China to make it appear that the US is the only good option and that the Philippines has no other choice but to side with it and it also justify its military presence there.
2014? - 2016 range also represents the evolution of the “Pivot to Asia Policy” into the “Indo-Pacific strategy”, perhaps the lack of major or significant protests beyond that represents the maturation and solidification of US influence in the Philippines, perhaps by using soft power to justify US presence there and prevent protests from happening. (Although, I don’t know what kind of control the US has on Philippine media outlets)
The US knows it can’t use too much propaganda or control in the Philippines otherwise the Philippines would get too suspicious and weary. So the US prioritizes stopping any Anti-US military sentiment over things like Pro-Palestine sentiment (I don’t think there is much the Philippines can do to impede Israel in the first place)
I think this was one of the moments the US used its control in the Philippines and why the “DoD sinovacc misinformation” news didn’t seem to become particularly popular in the Philippines. Because the Philippines would start asking questions like:
When I was checking Philippine media they seemed to be more concerned over China in the SCS, so this could be one of the US tactics, to deflect and distract attention.