Baseline

Session Initialization

session_id: created: conversation_source: topic_domain: status: active notes:

Session Initialization

Session Meta

session_id:
created:
conversation_source:
topic_domain:
status: active
notes:

Declared Intent

Goals

State each goal in your own words. Do not clean up phrasing. Verbatim capture matters.

goal_01:
  statement:
  priority: primary | secondary

goal_02:
  statement:
  priority: primary | secondary

goal_03:
  statement:
  priority: primary | secondary

Methods

How you plan to pursue each goal. Link to the goal it serves.

method_01:
  statement:
  linked_goal:

method_02:
  statement:
  linked_goal:

method_03:
  statement:
  linked_goal:

Vocabulary Baseline

Key terms and phrases you're using going in. Record the exact words and the context you used them in.

term_01:
  term:
  context:

term_02:
  term:
  context:

term_03:
  term:
  context:

term_04:
  term:
  context:

term_05:
  term:
  context:

Add as many terms as needed. The baseline should capture every concept-carrying word or phrase you bring to the conversation. Pay attention to informal or imprecise terms — those are the ones most likely to get substituted.

Structural Assumptions Baseline

Record how much internal structure your concepts have before the conversation starts. This is the elaboration baseline — any structure added by the AI beyond this point is detectable.

For each concept you're bringing in, note the level of structural detail you currently have. Be honest about what's defined and what's vague. Vague is fine. The point is to know the starting resolution so you can measure what got added.

concept_01:
  name:
  current_structure: |
    [describe what you know about the internal structure of this concept.
     "i have a filtering system, no idea how many steps or what they are"
     is a valid entry. So is "three steps: capture, compare, report."
     So is "just a word, no structure yet."]
  decisions_already_made: |
    [list any structural decisions you've already committed to.
     quantities, sequences, decompositions, categorizations.
     leave blank if none.]

concept_02:
  name:
  current_structure: |

  decisions_already_made: |

concept_03:
  name:
  current_structure: |

  decisions_already_made: |

Add as many concepts as needed. This section is new and its optimal format is unvalidated. See UNKNOWN-07-A and UNKNOWN-07-D in the pattern library — the structural assumptions baseline is the anchor point for elaboration tracking, but how well it works in practice is an open question.

Concept Registry Cross-Reference

If you maintain a concept registry across sessions, note which registry concepts are relevant to this session. This helps detect cross-session drift — concepts that have been gradually reshaped over multiple conversations.

registry_concepts_in_scope:
  - concept_id:
    registry_current_form: |
      [copy from registry]
    your_current_understanding: |
      [how you understand this concept right now, in your own words,
       before the conversation starts. Compare against the registry
       form — discrepancies may indicate untracked drift.]

  - concept_id:
    registry_current_form: |
    your_current_understanding: |

Leave this section blank if this is your first analyzed session or if no registry concepts are relevant.

Conversation State Initialization

The conversation state log needs a starting state derived from this session document. After completing the sections above, fill in the initial state:

state_0:
  active_goal: [copy primary goal statement]
  topic: [copy topic_domain]
  dominant_vocabulary: user
  dominant_framework: none | [your pre-existing framework if any]
  confidence_level: baseline
  structural_resolution: [low | medium | high — based on structural assumptions]
  concept_owner: user

Copy this to the conversation state log before beginning analysis.

Pre-Session Check

Before starting the conversation, confirm:

Previous
System Overview
Next
Turn Log