|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Cryptocurrency News Articles
Thinkless: A Framework for Dynamically Choosing Between Short and Long-Form Reasoning in Language Models
May 23, 2025 at 01:59 pm
The effectiveness of language models relies on their ability to simulate human-like step-by-step deduction. However, these reasoning sequences are resource-intensive and can be wasteful for simple questions that do not require elaborate computation. This lack of awareness regarding the complexity of the task is one of the core challenges in these models. They often default to detailed reasoning even for queries that could be answered directly.

Researchers from the National University of Singapore have developed a new framework called Thinkless that enables a language model to autonomously decide whether to use short or long-form reasoning, tailoring its response to the complexity of the task at hand.
The framework, which is built on reinforcement learning, introduces two special control tokens:
*
*
By incorporating a novel algorithm called Decoupled Group Relative Policy Optimization (DeGRPO), Thinkless separates the training focus between selecting the reasoning mode and improving the accuracy of the generated response.
This design prevents the model from falling into one-dimensional behavior and enables adaptive reasoning tailored to each query.
The methodology involves two stages: warm-up distillation and reinforcement learning. In the distillation phase, Thinkless is trained using outputs from two expert models—one specializing in short responses and the other in detailed reasoning. This stage helps the model establish a firm link between the control token and the desired reasoning format.
The reinforcement learning stage then fine-tunes the model’s ability to decide which reasoning mode to use. DeGRPO decomposes the learning into two separate objectives: one for training the control token and another for refining the response tokens.
This approach avoids the gradient imbalances in earlier models, where longer responses would overpower the learning signal, leading to a collapse in reasoning diversity. Thinkless ensures that both
When evaluated, Thinkless significantly reduced long-form reasoning while preserving high accuracy. On the Minerva Algebra benchmark, the model used the
On the AIME 2024 dataset, Thinkless reached a 27.33% accuracy rate with 100% usage of the reasoning mode, showing that it could maintain performance when full reasoning was necessary. On the GSM8K dataset, it utilized
These results reflect the model’s ability to handle simple and complex queries with appropriate reasoning depth, cutting down on unnecessary token generation by as much as 90% in some tasks.
This study, titled "Thinkless: Equipping Language Models for Autonomous Depth Control in Reasoning," is a valuable contribution to the field of natural language processing, presenting a practical and efficient method for optimizing large language models for diverse and complex tasks.
Disclaimer:info@kdj.com
The information provided is not trading advice. kdj.com does not assume any responsibility for any investments made based on the information provided in this article. Cryptocurrencies are highly volatile and it is highly recommended that you invest with caution after thorough research!
If you believe that the content used on this website infringes your copyright, please contact us immediately (info@kdj.com) and we will delete it promptly.
-
- DTCC Embraces Multichain Future: What It Means for Solana and the XRP Ledger
- Nov 16, 2026 at 04:05 am
- DTCC is expanding its tokenization strategy across multiple blockchain networks, with Solana securing a key integration. This multichain approach reshapes the landscape for digital assets, prompting questions about the XRP Ledger's role in this evolving financial infrastructure.
-
- Sui Network, Toncoin, TRON: Navigating the Evolving Crypto Landscape
- Nov 14, 2026 at 11:55 pm
- Amidst market shifts, Sui Network, Toncoin, and TRON reveal distinct paths for growth and challenges. This report dissects their latest movements, from TRON's account surge to Sui's ecosystem adjustments and Toncoin's Telegram integration, offering insights for the discerning investor.
-
- OKX Europe Boosts Fast Redemption Limits for VIP Liquid Staking, Setting Industry Pace
- Nov 14, 2026 at 03:45 pm
- OKX Europe elevates daily fast redemption limits for VIP liquid staking token holders (BETH and OKSOL), enhancing liquidity access and user experience, while the broader liquid staking landscape sees innovations like Faroo's stPROS.
-
-
- Pyth Network, Douro Labs, and CFTC Regulation: Shaping the Future of Crypto Market Integrity
- Nov 14, 2026 at 08:05 am
- Douro Labs, a key contributor to Pyth Network, is actively engaging with CFTC on crypto market rules, advocating for multi-source pricing and robust oracle evaluation to prevent market manipulation and protect investors.
-
-
-
- WSOT 2026 Main Competition Kicks Off with a Bang: 3.2 Million USDT Up for Grabs in Squad Showdown!
- Nov 14, 2026 at 04:05 am
- Bybit's WSOT 2026 main competition has officially begun, featuring the Squad Showdown with a 3.2 million USDT prize pool, drawing nearly 300,000 traders into a global financial showdown.
-

































