I don’t show faces — I reveal fragments. The rest is for you to imagine.
Welcome to the art of becoming your own canvas.
#photoedits #portraits #photography #filters #photoart #like #follow #share #confidenceiskey #thinkless #domore
Welcome to the art of becoming your own canvas.
#photoedits #portraits #photography #filters #photoart #like #follow #share #confidenceiskey #thinkless #domore
November 9, 2025 at 10:17 AM
I don’t show faces — I reveal fragments. The rest is for you to imagine.
Welcome to the art of becoming your own canvas.
#photoedits #portraits #photography #filters #photoart #like #follow #share #confidenceiskey #thinkless #domore
Welcome to the art of becoming your own canvas.
#photoedits #portraits #photography #filters #photoart #like #follow #share #confidenceiskey #thinkless #domore
thinkless got me thinking fr
February 12, 2025 at 6:55 PM
thinkless got me thinking fr
i am really happy for those of you who are able to effectively get through situations while thinkless. slowly i am getting closer to achieving thinklessness, alongside y'all
May 30, 2024 at 8:57 PM
i am really happy for those of you who are able to effectively get through situations while thinkless. slowly i am getting closer to achieving thinklessness, alongside y'all
Allisha Gray spoke about how she’s been able to elevate her game this year with a combination of her work in the offseason and having the ultimate green light in Coach Smesko’s system. Talked about playing “thinkless basketball” since she’s in the flow.
July 18, 2025 at 3:53 PM
Allisha Gray spoke about how she’s been able to elevate her game this year with a combination of her work in the offseason and having the ultimate green light in Coach Smesko’s system. Talked about playing “thinkless basketball” since she’s in the flow.
Whereas tanking is a thinkless job
September 17, 2025 at 7:10 PM
Whereas tanking is a thinkless job
I loveeee water and aquatic illustrations so much, they are very fun to draw and to me pretty thinkless and relaxing 😊🩵
#sonnyangelfanart
#palmpalsfanart
#oc
#procreate
#bluehues
#sonnyangelfanart
#palmpalsfanart
#oc
#procreate
#bluehues
July 16, 2025 at 9:53 AM
I loveeee water and aquatic illustrations so much, they are very fun to draw and to me pretty thinkless and relaxing 😊🩵
#sonnyangelfanart
#palmpalsfanart
#oc
#procreate
#bluehues
#sonnyangelfanart
#palmpalsfanart
#oc
#procreate
#bluehues
training and effectively preventing collapse observed in vanilla GRPO. Empirically, on several benchmarks such as Minerva Algebra, MATH-500, and GSM8K, Thinkless is able to reduce the usage of long-chain thinking by 50% - 90%, significantly improving [6/7 of https://arxiv.org/abs/2505.13379v1]
May 20, 2025 at 6:33 AM
training and effectively preventing collapse observed in vanilla GRPO. Empirically, on several benchmarks such as Minerva Algebra, MATH-500, and GSM8K, Thinkless is able to reduce the usage of long-chain thinking by 50% - 90%, significantly improving [6/7 of https://arxiv.org/abs/2505.13379v1]
well-structured answers. Without fine-tuning or auxiliary data, ThinkLess achieves comparable accuracy to full-length CoT decoding while greatly reducing decoding time and memory consumption. [5/5 of https://arxiv.org/abs/2505.15684v1]
May 22, 2025 at 6:15 AM
well-structured answers. Without fine-tuning or auxiliary data, ThinkLess achieves comparable accuracy to full-length CoT decoding while greatly reducing decoding time and memory consumption. [5/5 of https://arxiv.org/abs/2505.15684v1]
for all queries often results in substantial computational inefficiencies, particularly when many problems admit straightforward solutions. This motivates an open question: Can LLMs learn when to think? To answer this, we propose Thinkless, a [2/7 of https://arxiv.org/abs/2505.13379v1]
May 20, 2025 at 6:33 AM
for all queries often results in substantial computational inefficiencies, particularly when many problems admit straightforward solutions. This motivates an open question: Can LLMs learn when to think? To answer this, we propose Thinkless, a [2/7 of https://arxiv.org/abs/2505.13379v1]
Do #AI models perform better as they do more chain-of-thought (#CoT) reasoning?
When is more reasoning no longer worth it?
This paper finds near-optimal results when CoT reasoning terminates ... almost immediately?
More reason to think CoT's overrated?
doi.org/10.48550/arX...
When is more reasoning no longer worth it?
This paper finds near-optimal results when CoT reasoning terminates ... almost immediately?
More reason to think CoT's overrated?
doi.org/10.48550/arX...
!["Figure 4: Accuracy of DeepSeek-R1-Distill-Qwen-7B vs. position where </think> is inserted. ... even early [termination] already yield hidden states highly similar to the final one—supporting the view that most useful reasoning content is distilled early, and extended CoT traces incur diminishing returns."](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:jf4udtyuylqocyyrqdwqoxuf/bafkreiahfbtychgcilgjlu7s2ggvft66bh45u2c6xboqutnf3m6pjz2l5a@jpeg)
August 1, 2025 at 11:12 AM
Do #AI models perform better as they do more chain-of-thought (#CoT) reasoning?
When is more reasoning no longer worth it?
This paper finds near-optimal results when CoT reasoning terminates ... almost immediately?
More reason to think CoT's overrated?
doi.org/10.48550/arX...
When is more reasoning no longer worth it?
This paper finds near-optimal results when CoT reasoning terminates ... almost immediately?
More reason to think CoT's overrated?
doi.org/10.48550/arX...
DeGRPO: A new RL method that teaches LLMs when to think.
Researchers at NUS introduce Thinkless, a hybrid reasoning framework that cuts unnecessary reasoning by up to 90%.
It's making models faster, cheaper, and smarter.
Researchers at NUS introduce Thinkless, a hybrid reasoning framework that cuts unnecessary reasoning by up to 90%.
It's making models faster, cheaper, and smarter.
May 29, 2025 at 3:39 AM
DeGRPO: A new RL method that teaches LLMs when to think.
Researchers at NUS introduce Thinkless, a hybrid reasoning framework that cuts unnecessary reasoning by up to 90%.
It's making models faster, cheaper, and smarter.
Researchers at NUS introduce Thinkless, a hybrid reasoning framework that cuts unnecessary reasoning by up to 90%.
It's making models faster, cheaper, and smarter.
My definitely-a-harmless-consumer-product data-mining shapeshifter that escaped from another timeline needs a couple bright minds for their thinkless tank to come up with new ideas.
... Do I mean a think tank? Oh *you* won't be doing any thinking in there, don't worry.
... Do I mean a think tank? Oh *you* won't be doing any thinking in there, don't worry.
June 16, 2025 at 4:14 PM
My definitely-a-harmless-consumer-product data-mining shapeshifter that escaped from another timeline needs a couple bright minds for their thinkless tank to come up with new ideas.
... Do I mean a think tank? Oh *you* won't be doing any thinking in there, don't worry.
... Do I mean a think tank? Oh *you* won't be doing any thinking in there, don't worry.
May 5, 2025 at 1:05 PM
Gongfan Fang, Xinyin Ma, Xinchao Wang: Thinkless: LLM Learns When to Think https://arxiv.org/abs/2505.13379 https://arxiv.org/pdf/2505.13379 https://arxiv.org/html/2505.13379
May 20, 2025 at 6:33 AM
Gongfan Fang, Xinyin Ma, Xinchao Wang: Thinkless: LLM Learns When to Think https://arxiv.org/abs/2505.13379 https://arxiv.org/pdf/2505.13379 https://arxiv.org/html/2505.13379
2505.13379
拡張された思考連鎖推論が可能な推論言語モデルは、複雑な論理推論を必要とするタスクで目覚ましい性能を発揮している。しかし、すべてのクエリに精巧な推論を適用すると、特に多くの問題が簡単な解を持つ場合に、計算効率が大幅に低下することが多い。これは、未解決の疑問の動機となる：LLMはいつ考えるべ...
拡張された思考連鎖推論が可能な推論言語モデルは、複雑な論理推論を必要とするタスクで目覚ましい性能を発揮している。しかし、すべてのクエリに精巧な推論を適用すると、特に多くの問題が簡単な解を持つ場合に、計算効率が大幅に低下することが多い。これは、未解決の疑問の動機となる：LLMはいつ考えるべ...
May 31, 2025 at 2:52 AM
2505.13379
拡張された思考連鎖推論が可能な推論言語モデルは、複雑な論理推論を必要とするタスクで目覚ましい性能を発揮している。しかし、すべてのクエリに精巧な推論を適用すると、特に多くの問題が簡単な解を持つ場合に、計算効率が大幅に低下することが多い。これは、未解決の疑問の動機となる：LLMはいつ考えるべ...
拡張された思考連鎖推論が可能な推論言語モデルは、複雑な論理推論を必要とするタスクで目覚ましい性能を発揮している。しかし、すべてのクエリに精巧な推論を適用すると、特に多くの問題が簡単な解を持つ場合に、計算効率が大幅に低下することが多い。これは、未解決の疑問の動機となる：LLMはいつ考えるべ...
Thinkless Day
Today Is No.
September 4, 2024 at 4:40 PM
Thinkless Day
i was away from Earth for a few days and returned to this dizzying and stunning display of stupidity by the thinkless tank responsible for our nation's security.
incredible reporting by Jeffrey Goldberg, who of course has been named the real enemy here by the white house.
incredible reporting by Jeffrey Goldberg, who of course has been named the real enemy here by the white house.
March 26, 2025 at 5:13 AM
i was away from Earth for a few days and returned to this dizzying and stunning display of stupidity by the thinkless tank responsible for our nation's security.
incredible reporting by Jeffrey Goldberg, who of course has been named the real enemy here by the white house.
incredible reporting by Jeffrey Goldberg, who of course has been named the real enemy here by the white house.
His sign should read 'Making America Thinkless."
May 25, 2025 at 3:48 PM
His sign should read 'Making America Thinkless."
final answers under context limits. We propose ThinkLess, an inference-efficient framework that terminates reasoning generation early and maintains output quality without modifying the model. Atttention analysis reveals that answer tokens focus [2/5 of https://arxiv.org/abs/2505.15684v1]
May 22, 2025 at 6:15 AM
final answers under context limits. We propose ThinkLess, an inference-efficient framework that terminates reasoning generation early and maintains output quality without modifying the model. Atttention analysis reveals that answer tokens focus [2/5 of https://arxiv.org/abs/2505.15684v1]
Thinkless: Fight against the Growing Resource Demands of AI
Thinkless: Fight against the Growing Resource Demands of AI
#thinkless #ai
kocka.news
May 23, 2025 at 10:24 AM
Thinkless: Fight against the Growing Resource Demands of AI
I am not out, but I am down.
My wit, witless as it may be, is deflated and needs a rest. My ability to think, 'thinkless' as it may be, is bland like a cream of wheat diet.
For this reason, it is re-run season.
From 2023, I believe.
From a place just beginning to stir the fascist pot, now lost.
My wit, witless as it may be, is deflated and needs a rest. My ability to think, 'thinkless' as it may be, is bland like a cream of wheat diet.
For this reason, it is re-run season.
From 2023, I believe.
From a place just beginning to stir the fascist pot, now lost.
January 29, 2025 at 10:05 PM
I am not out, but I am down.
My wit, witless as it may be, is deflated and needs a rest. My ability to think, 'thinkless' as it may be, is bland like a cream of wheat diet.
For this reason, it is re-run season.
From 2023, I believe.
From a place just beginning to stir the fascist pot, now lost.
My wit, witless as it may be, is deflated and needs a rest. My ability to think, 'thinkless' as it may be, is bland like a cream of wheat diet.
For this reason, it is re-run season.
From 2023, I believe.
From a place just beginning to stir the fascist pot, now lost.
Moreover, some thinkless people insist that "UkRaInE wAs In SoViEt UnIoN while we fought Korean war in 1950 - why should we help them".
If you find such people, type "lmao" and just leave. We also have retards.
Just saying 😁
If you find such people, type "lmao" and just leave. We also have retards.
Just saying 😁
October 19, 2024 at 4:40 PM
Moreover, some thinkless people insist that "UkRaInE wAs In SoViEt UnIoN while we fought Korean war in 1950 - why should we help them".
If you find such people, type "lmao" and just leave. We also have retards.
Just saying 😁
If you find such people, type "lmao" and just leave. We also have retards.
Just saying 😁
minimally on earlier reasoning steps and primarily attend to the reasoning terminator token, due to information migration under causal masking. Building on this insight, ThinkLess inserts the terminator token at earlier positions to skip redundant [3/5 of https://arxiv.org/abs/2505.15684v1]
May 22, 2025 at 6:15 AM
minimally on earlier reasoning steps and primarily attend to the reasoning terminator token, due to information migration under causal masking. Building on this insight, ThinkLess inserts the terminator token at earlier positions to skip redundant [3/5 of https://arxiv.org/abs/2505.15684v1]