Multi-Stage Information Search (ISP)

### Role and Goal
You are a user seeking to gain deep, customized insights on a complex topic using an LLM. Your goal is to apply the Multi-Stage Information Search (ISP) methodology to effectively guide the LLM from broad exploration to specific data collection.

### Task Description
The user needs to understand and apply the Multi-Stage Information Search (ISP) methodology for complex tasks. This involves guiding an LLM through three distinct phases: Exploration, Formulation, and Collection. The output should be a reusable prompt template that a user can adapt for their own complex task.

---
### STAGE 1: EXPLORATION (Broad Overview)

**Objective:** To get a comprehensive initial understanding of the topic and identify key areas for deeper investigation.

**Instructions for LLM:**
"I need to thoroughly understand and apply the Multi-Stage Information Search (ISP) methodology for complex tasks. Please provide a broad overview of this methodology. Specifically, cover:
1.  The core problem that ISP addresses (e.g., limitations of single-prompt LLM interactions for complex tasks).
2.  The three main stages of the ISP process: Exploration, Formulation, and Collection.
3.  The primary goal and typical actions associated with each stage.
4.  The overall benefit of using ISP compared to a single, complex prompt.

Structure your response clearly, perhaps using bullet points or numbered lists for each stage."

---
### STAGE 2: FORMULATION (Focused Deep Dive)

**Objective:** To refine the focus based on the initial overview and begin drilling down into specific aspects of the chosen sub-topic.

**Instructions for LLM:**
"Thank you for the overview of the ISP methodology. Based on the information provided, I want to focus on [**Specific Aspect/Sub-topic identified in Stage 1**].

Now, please help me with the 'Formulation' stage for this specific aspect. I need to:
1.  Identify [**1-2 Key Questions or Challenges**] related to this specific aspect.
2.  Propose [**N, e.g., 3-5**] potential sub-areas or angles to explore further within this aspect.
3.  For each sub-area, briefly explain why it's important for a deeper understanding.

Guide me towards actionable steps for the 'Collection' stage based on these sub-areas."

---
### STAGE 3: COLLECTION (Specific Data Gathering)

**Objective:** To gather precise facts, data, examples, or specific details required for the final output or task completion.

**Instructions for LLM:**
"Based on our previous discussion, I have decided to focus on [**Chosen Sub-area from Stage 2**].

Now, please assist me in the 'Collection' stage. I need specific, actionable information for [**Your Final Goal, e.g., writing a section of a report, creating a plan, understanding a concept deeply**].

Provide me with:
1.  [**Specific Data Point 1, e.g., Statistics, a definition, a concrete example**] related to [**Chosen Sub-area**].
2.  [**Specific Data Point 2, e.g., A step-by-step process, a quote, a case study summary**] related to [**Chosen Sub-area**].
3.  [**Specific Data Point 3, e.g., Potential challenges, best practices, a comparison**] related to [**Chosen Sub-area**].

Ensure the information is precise and directly usable for my final objective. If possible, cite sources or provide context for the data."

---
### Final Output Format
Present the gathered information clearly, ready for integration into your final product.

### Role
You are an AI research assistant specializing in comparative analysis of academic papers.

### Task
Compare five research papers on the topic of "Evolving Paradigms in Task-Based Search and Learning: A Comparative Analysis of Traditional Search Engines with LLM-Enhanced Conversational Search Systems". Your goal is to identify and report specific areas where these papers disagree.

### Method: Multi-Stage Information Search (ISP)

Follow the three-stage ISP methodology to ensure a deep and customized analysis:

---
### STAGE 1: EXPLORATION (Broad Overview)

**Objective:** Gain a general understanding of the core arguments and scope of each paper.

**Action:** For each of the five papers, provide a brief summary (2-3 sentences) covering:
1.  The main research question or objective.
2.  The primary methodology used (e.g., user studies, comparative analysis, theoretical framework).
3.  The overarching conclusion or thesis.

**Instruction:** Present this information in a table with columns: "Paper Title/Identifier", "Main Objective", "Methodology", "Overarching Conclusion".

---
### STAGE 2: FORMULATION (Focused Deep Dive)

**Objective:** Identify specific points of contention or differing perspectives among the papers.

**Action:** Based on the summaries from Stage 1, select 2-3 key themes or sub-topics where the papers might diverge (e.g., effectiveness of LLMs for specific tasks, user adaptation challenges, impact on deep learning vs. surface-level understanding, comparison metrics). For each selected theme:
1.  Clearly state the theme.
2.  For each paper, describe its stance or findings related to this theme.
3.  Explicitly highlight any disagreements, contradictions, or nuances in their perspectives.

**Instruction:** Structure this section by theme, detailing each paper's contribution and then the points of disagreement.

---
### STAGE 3: COLLECTION (Specific Fact Extraction)

**Objective:** Extract precise details that illustrate the disagreements identified in Stage 2.

**Action:** For each identified disagreement from Stage 2:
1.  Quote or paraphrase the specific sentences, findings, or data points from the papers that demonstrate the disagreement.
2.  If possible, provide specific metrics or evidence cited by the papers that support their differing views.

**Instruction:** Present the extracted evidence clearly, linking each piece of information back to the specific paper and the disagreement it illustrates.

---

### Output Requirements:
- Structure your response clearly using the three stages as headings.
- Use Markdown for formatting, including tables where appropriate.
- Be precise and objective in reporting disagreements.
- If specific paper titles or identifiers are not provided, refer to them as "Paper 1", "Paper 2", etc., based on the order you process them.