Scribd Logo
Upload

Welcome to Scribd!

  • 12 views

    Chain of Thought

    Uploaded by

    nihalgupta339
    Learn chain of thoughts

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Chain of Thought

    Uploaded by

    nihalgupta339
    12 views4 pages
    Learn chain of thoughts

    Original Title

    CHAIN OF THOUGHT

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Learn chain of thoughts

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    12 views4 pages

    Chain of Thought

    Uploaded by

    nihalgupta339
    Learn chain of thoughts

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 4

    Chain-of-Thought (CoT) Prompting

    Introduction of intermediate “reasoning” steps, improves the


    performance of LLMs in tasks that require complex reasoning
    like arithmetic, common sense, and symbolic reasoning.

    In their paper, “Chain-of-Thought Prompting Elicits Reasoning


    in Large Language Models”, Wei et. al. demonstrated how
    LLMs naturally start reasoning with a few examples.

    In this technique, a few logical reasoning steps are added to


    the prompt as examples for the LLM to understand how to
    arrive at the desired outcome.

    Wei et al. (2022)


    Chain-of-Thought Prompting Elicits Reasoning in Large Language Models

    Another idea of “Zero Shot CoT” was introduced by


    Kojima et al. 2022 where, instead of adding examples
    for Few Shot CoT, we just add “Let’s think step by
    step” to the prompt.
    abhinav-kimothi
    Automatic Chain-of-Thought (Auto-CoT)
    As we saw, CoT prompting involves creating examples for the
    LLM. This is a manual process and introduces subjectivity. To
    reduce this subjectivity, Zhang et al. (2022) introduced Auto-
    COT. There are two stages involved in Auto-CoT

    A: Question Clustering
    A dataset of diverse questions Questions clustered into ‘k’ groups

    1 3 k-1

    2 4 k

    Stage A: Create clusters from a dataset of diverse question

    B: Demonstration Sampling

    Demo Selecting one representative


    question from each cluster
    Q: While shopping for
    music online....
    A : Let’s think step by

    LLM
    step....
    ...
    Q:
    A:..

    Reasoning Chain for Zero Shot CoT by adding “Let’s


    representative questions think step by step”
    Stage B: Select one question from each cluster and generate its reasoning chain
    using Zero-Shot-CoT with simple heuristics

    The Questions with Reasoning Chain in the Demo are then


    used as examples for new questions

    Checkout Auto-CoT Code


    abhinav-kimothi
    Benefits of Chain of Thought Prompting

    Breaks down multi-step problems into simpler


    components to enable more efficient solving

    Provides transparency into models' reasoning


    for interpretability

    Applicable across diverse reasoning tasks like


    math, commonsense, and symbolic
    manipulation.

    Easily integrated into existing models via


    prompting. Does not require any architectural
    change

    Makes models' thought processes relatable to


    facilitate human-AI collaboration

    Adapts complexity of reasoning chain to task


    difficulty for broad applicability

    Enables error identification by exposing


    models' step-by-step reasoning logic

    Teaches generalizable structured problem-


    solving strategies transferable across tasks.

    abhinav-kimothi
    Limitations of CoT

    Task Complexity Prompt Quality


    Chain of Thought Prompting The technique depends heavily
    offers minimal additional value on prompt quality to steer
    over standard prompting for tasks models through reasoning
    that lack multi-step reasoning chains. Crafting prompts that
    requirements or cannot be easily provide effective stepwise
    decomposed. Its benefits are best guidance demands care and can
    achieved for problems requiring prove difficult for complex
    sequential logic or intermediate domains necessitating expert
    explanatory steps knowledge.

    Scalability Model Size


    While Auto CoT tries to automate Chain of Thought reasoning works
    the process of creating the well only on very large models
    reasoning chains, it still remains a with more than 100 Billion
    complex and labor intensive parameters. On smaller models
    process to create them. As the the efficiency reduces. On the
    tasks increase, the manual effort other hand, the efficacy of CoT
    in creating or verifying the remains to be seen as the model
    reasoning chains keeps on size increases further.
    increasing.

    abhinav-kimothi

    You might also like