Prompt engineering for text generation: Difference between revisions

← Older edit

Prompt engineering for text generation (view source)

Revision as of 07:32, 16 January 2024

6,481 bytes added , 16 January

no edit summary

Beetlejuice

1,065

edits

@@ Line 140: / Line 140: @@
 ====Embeddings via Contrastive Learning====
-Rubin et al. (2022) suggested training embeddings through [[contrastive learning]] specific to one [[training dataset]] for in-context learning sample selection. This approach measures the quality of an example based on a conditioned probability assigned by the language model.<ref name="”114”">Rubin et al. (2022) Learning To Retrieve Prompts for In-Context Learning https://arxiv.org/abs/2112.08633</ref>
+Rubin et al. (2022) suggested training embeddings through [[contrastive learning]] specific to one [[training dataset]] for [[in-context learning]] sample selection. This approach measures the quality of an example based on a conditioned probability assigned by the language model.<ref name="”114”">Rubin et al. (2022) Learning To Retrieve Prompts for In-Context Learning https://arxiv.org/abs/2112.08633</ref>
 ====Q-Learning====
@@ Line 149: / Line 149: @@
 ===Tips for Example Ordering===
-A general recommendation is to maintain a diverse selection of examples relevant to the test sample and present them in random order to avoid majority label bias and recency bias. Increasing model sizes or including more training examples does not necessarily reduce variance among different permutations of in-context examples. The exact order may work well for one model but poorly for another.
+A general recommendation is to maintain a diverse selection of examples relevant to the test sample and present them in random order to avoid [[majority label bias]] and [[recency bias]]. Increasing [[model size]]s or including more [[training examples]] does not necessarily reduce [[variance]] among different permutations of in-context examples. The exact order may work well for one model but poorly for another.
-When the validation set is limited, Lu et al. (2022) suggested choosing the order such that the model does not produce extremely unbalanced predictions or exhibit overconfidence in its predictions.<ref name="”117”">Lu et al. (2022) Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivity https://arxiv.org/abs/2104.08786</ref>
+When the [[validation set]] is limited, Lu et al. (2022) suggested choosing the order such that the model does not produce extremely unbalanced predictions or exhibit overconfidence in its predictions.<ref name="”117”">Lu et al. (2022) Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivity https://arxiv.org/abs/2104.08786</ref>
 ==Roles==
@@ Line 201: / Line 201: @@
 *'''[[Straightforward]]''' and '''[[Professional]]''' - business emails, formal communication, legal documents
 *'''[[Trustworthy]]''' and '''[[Professional]]''' - business proposals, executive summaries, investor pitches
+==Self-Consistency Sampling==
+[[Self-consistency sampling]] is a method for generating multiple outputs using a [[temperature]] greater than 0 and selecting the best candidate from the generated outputs. The criteria for choosing the best candidate may vary according to the task. A common approach is to use [[majority vote]]. In tasks that are easy to validate, such as programming questions with unit tests, the outputs can be run through an interpreter and their correctness can be verified using unit tests.<ref name="”118”">Wang et al. (2022a) Self-Consistency Improves Chain of Thought Reasoning in Language Models https://arxiv.org/abs/2203.11171</ref>
 ==Chain of Thought Prompting==
-{{see also|Chain of Thought Prompting}}
+[[Chain of Thought Prompting]] (CoT prompting) is a technique introduced by Wei et al. (2022) to generate a sequence of short sentences describing step-by-step reasoning, known as [[reasoning chains]] or [[rationales]], leading to the final answer. [[CoT prompting]] is particularly useful for complex reasoning tasks when applied to large language models (e.g., those with over 50 billion parameters), while simpler tasks may benefit only marginally.<ref name="”119”">Wei et al. (2022) Chain-of-Thought Prompting Elicits Reasoning in Large Language Models https://arxiv.org/abs/2201.11903</ref>
+===Types of CoT Prompts===
+There are two main types of CoT prompting:
+====Few-shot CoT====
+[[Few-shot CoT prompting]] involves providing the model with a limited number of demonstrations, each containing either manually written or model-generated high-quality reasoning chains. Examples of such demonstrations are provided in the original article, showcasing how this type of prompting is used to solve various mathematical reasoning problems.
+====Zero-shot CoT====
+[[Zero-shot CoT prompting]] uses natural language statements, such as "Let's think step by step" or "Let's work this out step by step to be sure we have the right answer," to explicitly encourage the model to generate reasoning chains. Following this, a statement like "Therefore, the answer is" is used to prompt the model to produce the final answer.<ref name="”128”">Kojima et al. (2022) Large Language Models are Zero-Shot Reasoners https://arxiv.org/abs/2205.11916</ref><ref name="”129”">Zhou et al. (2022) Large Language Models Are Human-Level Prompt Engineers https://arxiv.org/abs/2211.01910</ref>
+===Tips and Extensions===
+Several techniques have been proposed to improve the accuracy and effectiveness of CoT prompting:
+*[[Self-consistency sampling]], as suggested by Wang et al. (2022a), can improve reasoning accuracy by sampling a number of diverse answers and taking the majority vote.<ref name="”118”"></ref>
+*Wang et al. (2022b) proposed using ensemble learning by altering the example order or replacing human-written rationales with model-generated ones, introducing randomness during multiple sample trials. Model outputs can then be aggregated using a majority vote to obtain the final answer.<ref name="”120”">Wang et al. (2022b) Rationale-Augmented Ensembles in Language Models https://arxiv.org/abs/2207.00747</ref>
+*If training examples only have true answers but no rationales, the [[STaR]] ([[Self-Taught Reasoner]]) method by Zelikman et al. (2022) can be followed: (1) ask the model to generate reasoning chains and keep only those leading to correct answers; (2) fine-tune the model with generated rationales and repeat the process until convergence. Higher temperature settings are more likely to generate incorrect rationales with correct answers.<ref name="”121”">Zelikman et al. (2022) STaR: Bootstrapping Reasoning With Reasoning https://arxiv.org/abs/2203.14465</ref>
+*Fu et al. (2023) found that prompts with demonstrations of higher reasoning complexity lead to better performance. They also suggested that using newline (\n) symbols to separate reasoning steps works better than step indicators, periods, or semicolons.<ref name="”122”">Fu et al. (2023) Complexity-Based Prompting for Multi-Step Reasoning https://arxiv.org/abs/2210.00720</ref>
+*Complexity-based consistency, as proposed by Fu et al. (2023), involves explicitly preferring complex chains among all generations by taking a majority vote among only the top complex chains.<ref name="”122”"></ref>
+*Shum et al. (2023) discovered that CoT prompts with only complex examples improve the accuracy of complex questions but perform poorly on simple questions. This finding was based on evidence from the [[GSM8k]] dataset.<ref name="”123”">Shum et al. (2023) Automatic Prompt Augmentation and Selection with Chain-of-Thought from Labeled Data https://arxiv.org/abs/2302.12822</ref>
+*Fu et al. (2023) found that changing "Q:" to "Question:" in the prompts is helpful.<ref name="”122”"></ref>
+*Ye & Durrett (2022) observed that including explanations in prompts has a small to moderate effect on [[NLP]] tasks that involve reasoning over text, such as [[question-answering]] (QA) and [[natural language inference]] (NLI). They also noted that nonfactual explanations are more likely to lead to incorrect predictions than inconsistent explanations.<ref name="”124”">Ye & Durrett (2022) The Unreliability of Explanations in Few-shot Prompting for Textual Reasoning https://arxiv.org/abs/2205.03401</ref>
+*[[Self-Ask]], a method proposed by Press et al. (2022), repeatedly prompts the model to ask follow-up questions, constructing the thought process iteratively.<ref name="”125”">Press et al. (2022) Measuring and Narrowing the Compositionality Gap in Language Models https://arxiv.org/abs/2210.03350</ref> Search engine results can be used to answer these follow-up questions. Similarly, [[IRCoT]] ([[Interleaving Retrieval CoT]]; Trivedi et al. 2022) and [[ReAct]] ([[Reason + Act]]; Yao et al. 2023) combine iterative CoT prompting with queries to Wikipedia APIs. These methods search for relevant entities and content and then incorporate the retrieved information back into the context, further enhancing the model's reasoning capabilities.<ref name="”126”">Trivedi et al. (2022) Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions https://arxiv.org/abs/2212.10509</ref><ref name="”127”">Yao et al. (2023) ReAct: Synergizing Reasoning and Acting in Language Models https://arxiv.org/abs/2210.03629</ref>
+==26 Principals for Good Prompts==
+{{see also|26 Principles of Good Prompts}}
+{{:26 Principles of Good Prompts}}
 ==Prompt Engineering for Code Generation Models==
@@ Line 281: / Line 318: @@
 *[[Vividness]] -
 *[[Ecclesiastical]] -
+==Connecting External APIs==
 ==Resources==