datamountaincmcastelli

I ran a quick experiment investigating how DeepSeek-R1 performs on agentic jobs, in spite of not supporting tool usage natively, and I was quite amazed by initial outcomes. This experiment runs DeepSeek-R1 in a single-agent setup, where the model not only prepares the actions but likewise formulates the actions as executable Python code. On a subset1 of the GAIA recognition split, DeepSeek-R1 outshines Claude 3.5 Sonnet by 12.5% absolute, wiki.eqoarevival.com from 53.1% to 65.6% correct, and other models by an even larger margin:

The experiment followed model use guidelines from the DeepSeek-R1 paper and the design card: Don't utilize few-shot examples, avoid adding a system prompt, and set the temperature level to 0.5 - 0.7 (0.6 was utilized). You can find further assessment details here.

Approach

DeepSeek-R1's strong coding abilities enable it to function as a representative without being clearly trained for tool use. By permitting the model to produce actions as Python code, it can flexibly interact with environments through code execution.

Tools are carried out as Python code that is included straight in the prompt. This can be an easy function meaning or a module of a bigger package - any legitimate Python code. The model then generates code actions that call these tools.

Arise from executing these actions feed back to the design as follow-up messages, driving the next steps until a last response is reached. The agent framework is an easy iterative coding loop that moderates the conversation in between the design and its environment.

Conversations

DeepSeek-R1 is used as chat model in my experiment, where the design autonomously pulls additional context from its environment by utilizing tools e.g. by utilizing a search engine or bring information from web pages. This drives the conversation with the environment that continues till a final response is reached.

In contrast, o1 models are known to carry out badly when utilized as chat designs i.e. they do not try to pull context during a conversation. According to the connected short article, o1 designs carry out best when they have the full context available, with clear guidelines on what to do with it.

Initially, king-wifi.win I likewise attempted a complete context in a single timely technique at each action (with outcomes from previous steps consisted of), however this led to substantially lower scores on the GAIA subset. Switching to the conversational method explained above, I was able to reach the reported 65.6% performance.

This raises an intriguing concern about the claim that o1 isn't a chat model - perhaps this observation was more relevant to older o1 designs that lacked tool usage abilities? After all, isn't tool use support an important system for enabling models to pull additional context from their environment? This conversational approach certainly appears efficient for DeepSeek-R1, though I still require to perform comparable experiments with o1 designs.

Generalization

Although DeepSeek-R1 was mainly trained with RL on math and coding jobs, it is exceptional that generalization to agentic tasks with tool usage by means of code actions works so well. This ability to generalize to agentic tasks reminds of recent research by DeepMind that reveals that RL generalizes whereas SFT remembers, although generalization to tool use wasn't examined in that work.

Despite its capability to generalize to tool usage, DeepSeek-R1 frequently produces long thinking traces at each action, compared to other designs in my experiments, restricting the effectiveness of this model in a single-agent setup. Even simpler tasks often take a long time to complete. Further RL on agentic tool use, be it through code actions or not, could be one option to improve efficiency.

Underthinking

I likewise observed the underthinking phenomon with DeepSeek-R1. This is when a reasoning design regularly switches in between different reasoning thoughts without adequately exploring appealing paths to reach a correct service. This was a major reason for extremely long reasoning traces produced by DeepSeek-R1. This can be seen in the recorded traces that are available for download.

Future experiments

Another common application of thinking models is to use them for planning only, while using other models for creating code actions. This might be a potential brand-new function of freeact, if this separation of functions shows helpful for more complex tasks.

I'm likewise curious about how thinking models that currently usage (like o1, o3, ...) perform in a single-agent setup, with and without creating code actions. Recent advancements like OpenAI's Deep Research or Hugging Face's open-source Deep Research, which likewise uses code actions, look fascinating.