5). Handling context length Remember that reasoning tokens count as output tokens and affect your context limit. OpenAI recommends at least 25K tokens for reasoning and outputs when starting, then adjust as needed. Outputs that exceed the model’s context window are truncated.
@dair_ai
-
Reasoning Models: Provide More Details When Iterating
By
–
6). Provide more details as you iterate When working with reasoning models, start with a clear problem statement. Unlike traditional LLMs, reasoning models benefit from explicit constraints and limitations rather than step-by-step guidance.
-
Systematic Survey of Automatic Prompt Optimization Methods
By
–
9). A Systematic Survey of Prompt Optimization This paper offers a comprehensive survey of Automatic Prompt Optimization (APO)—defining its scope, presenting a unifying 5-part framework, categorizing existing methods,…
-
Protein LLMs: Architectures, Training, and Applications Overview
By
–
10). Protein LLMs A comprehensive overview of Protein LLMs, including architectures, training datasets, evaluation metrics, and applications.
-
LightThinker: Dynamic Compression of LLM Reasoning Steps
By
–
8). LightThinker This new paper proposes a novel approach to dynamically compress reasoning steps in LLMs, significantly improving efficiency without sacrificing accuracy.
-
PlanGEN: Multi-Agent Framework Enhances LLM Planning
By
–
6). PlanGEN PlanGEN is a multi-agent framework designed to enhance planning and reasoning in LLMs through constraint-guided iterative verification and adaptive algorithm selection.
-
METAL Multi-Agent Framework Enhances Automatic Chart-to-Code Generation
By
–
7). A Multi-Agent Framework for Chart Generation METAL is a vision-language model (VLM)-based multi-agent framework designed to significantly enhance automatic chart-to-code generation by decomposing the task into specialized iterative steps.
-
OpenAI Introduces GPT-4.5 with Enhanced Safety and Alignment
By
–
2). GPT-4.5 OpenAI introduces GPT-4.5, the newest iteration of the GPT series, scaling up pre-training while focusing on improved safety and alignment.
-
Chain-of-Draft Reduces Latency in Reasoning Language Models
By
–
3). Chain-of-Draft To address the issue of latency in reasoning LLMs, this work introduces Chain-of-Draft (CoD).
-
Anthropic releases Claude 3.7 Sonnet with extended thinking mode
By
–
1). Claude 3.7 Sonnet Anthropic releases a system card for its latest hybrid reasoning model, Claude 3.7 Sonnet, detailing safety measures, evaluations, and a new "extended thinking" mode.