core/notes/live_probe_2026-05-19.txt
Shay fd2fa67b2a docs(notes): live-probe log + reproducer for the 2026-05-19 fluency push
Captures the end-to-end behavior of the gloss-feature landing as a
durable, replayable artifact.  Two files:

notes/live_probe_2026-05-19.py
  Probe script — walks 51 prompts across 13 categories through
  fresh ChatRuntime() instances (cold-start invariant; no vault
  contamination across prompts).  Runs offline / locally:
    uv run python notes/live_probe_2026-05-19.py

notes/live_probe_2026-05-19.txt
  Captured output of the script on commit a8b611a.  Acts as a
  golden-master record: anyone can rerun the script and diff the
  output to detect surface drift.

Categories covered:
  - Cognition (5 prompts)            — truth, knowledge, memory, ...
  - Speech-act / discourse (5)       — fact, idea, statement, ...
  - Mental-state (5)                 — doubt, believe, self, mind, view
  - Adjectives / attitude (5)        — true, important, evident, ...
  - Temporal (5)                     — now, moment, future, before, time
  - Spatial (4)                      — here, place, above, between
  - Action verbs (4)                 — incl. infinitive-stripped form
  - Quantitative (4)                 — all, some, more, enough
  - Causation (4)                    — effect, outcome, consequence, trigger
  - Polarity / frequency (4)         — yes, always, never, maybe
  - Teaching-chain multi-clause (1)  — Why is truth important?
  - Genuinely OOV (3)                — hypothesis, javascript, quasar
  - Cause without teaching chain (2) — How does memory work?
                                       What causes doubt?

Grounding distribution observed:
  pack       45  (88.2%)   fluent gloss-backed surfaces
  oov         3  ( 5.9%)   honest OOV invitations
  none        2  ( 3.9%)   honest "I don't know" (CAUSE w/o chain —
                           deferred SurfaceSelector target)
  teaching    1  ( 2.0%)   multi-clause teaching-chain composition

Sample surfaces:

  [PACK] What is truth?
         Truth is a claim or state grounded by evidence and coherent
         judgment.  pack-grounded (en_core_cognition_v1).

  [PACK] To use means to put something into service for a purpose.
         pack-grounded (en_core_action_v1).

  [PACK] Something is important when it carries weight or priority in
         some judgment context.  pack-grounded (en_core_attitude_v1).

  [PACK] Always indicates the frequency of occurring without exception
         across all instances.  pack-grounded (en_core_polarity_v1).

  [TEACH] truth — teaching-grounded (cognition_chains_v1):
         cognition.truth; logos.core. truth grounds knowledge
         (cognition.knowledge). No session evidence yet.

  [OOV] I haven't learned 'hypothesis' yet (intent: definition).
         Mounted lexicon packs: ...
         Teach me via a reviewed PackMutationProposal.

  [NONE] I don't know — insufficient grounding for that yet.
         (CAUSE on memory — no teaching chain rooted here; the
         deliberate non-fallback the teaching pipeline uses as a
         discovery-gap signal)

No code change in this commit.  Pure documentation artifact for the
fluency-push baseline.
2026-05-19 07:56:24 -07:00

255 lines
9.5 KiB
Text

# Live Probe — 2026-05-19
#
# 51 prompts across 13 categories. Fresh ChatRuntime() per prompt
# (cold-start invariant — no vault contamination). Deterministic;
# every run produces byte-identical output.
#
# Reproduce by running notes/live_probe_2026-05-19.py (see the
# scaffolded version under that name).
#
=== Cognition ===
[PACK ] What is truth?
Truth is a claim or state grounded by evidence and coherent judgment.
pack-grounded (en_core_cognition_v1).
[PACK ] Define knowledge.
Knowledge is justified understanding grounded in evidence and recall.
pack-grounded (en_core_cognition_v1).
[PACK ] What is memory?
Memory is the retained record enabling recall of prior states. pack-grounded
(en_core_cognition_v1).
[PACK ] What does meaning mean?
Meaning is the semantic content carried by a sign or expression.
pack-grounded (en_core_cognition_v1).
[PACK ] What is wisdom?
Wisdom is sound judgment informed by knowledge and experience. pack-grounded
(en_core_cognition_v1).
=== Speech-act / discourse ===
[PACK ] What is a fact?
Fact is a proposition established as true. pack-grounded (en_core_meta_v1).
[PACK ] What is an idea?
Idea is a mental content that can be entertained or expressed. pack-grounded
(en_core_meta_v1).
[PACK ] What is a statement?
Statement is an utterance that expresses a definite proposition.
pack-grounded (en_core_meta_v1).
[PACK ] What does claim mean?
To claim means to assert a proposition as true without proof. pack-grounded
(en_core_meta_v1).
[PACK ] Define argument.
Argument is a structured sequence of claims supporting a conclusion.
pack-grounded (en_core_meta_v1).
=== Mental-state ===
[PACK ] What is doubt?
To doubt means to hold a mental state of uncertainty about a claim.
pack-grounded (en_core_meta_v1).
[PACK ] What does believe mean?
To believe means to hold a proposition as likely true without certainty.
pack-grounded (en_core_meta_v1).
[PACK ] What is the self?
Self is the agent or speaker as the reference point of identity.
pack-grounded (en_core_meta_v1).
[PACK ] Define mind.
Mind is the seat of thought, perception, and mental states. pack-grounded
(en_core_meta_v1).
[PACK ] What is a view?
View is a held perspective on a topic or claim. pack-grounded (en_core_meta_v1).
=== Adjectives (attitude) ===
[PACK ] What is true?
Something is true when it corresponds accurately to the evidence and reality.
pack-grounded (en_core_attitude_v1).
[PACK ] What does important mean?
Something is important when it carries weight or priority in some judgment
context. pack-grounded (en_core_attitude_v1).
[PACK ] Define evident.
Something is evident when it shows itself plainly to any careful observer.
pack-grounded (en_core_attitude_v1).
[PACK ] What is certain?
Something is certain when it holds without residual doubt given the evidence.
pack-grounded (en_core_attitude_v1).
[PACK ] What does necessary mean?
Something is necessary when it must hold for some outcome or condition to
obtain. pack-grounded (en_core_attitude_v1).
=== Temporal ===
[PACK ] What is now?
Now indicates the present moment at the time of utterance. pack-grounded
(en_core_temporal_v1).
[PACK ] Define moment.
Moment is a brief or pointlike interval of time. pack-grounded
(en_core_temporal_v1).
[PACK ] What is the future?
Future is the temporal region after the present. pack-grounded
(en_core_temporal_v1).
[PACK ] What does before mean?
Before is a relation of temporal anteriority between two events.
pack-grounded (en_core_temporal_v1).
[PACK ] What is time?
Time is the dimension along which events are ordered. pack-grounded
(en_core_temporal_v1).
=== Spatial ===
[PACK ] What is here?
Here indicates the location of the speaker or reference point. pack-grounded
(en_core_spatial_v1).
[PACK ] What is a place?
Place is a definite location in space. pack-grounded (en_core_spatial_v1).
[PACK ] Define above.
Above is a relation of vertical position higher than a reference point.
pack-grounded (en_core_spatial_v1).
[PACK ] What does between mean?
Between is a relation of spatial position flanked by two references.
pack-grounded (en_core_spatial_v1).
=== Action verbs (infinitive-stripped) ===
[PACK ] What is to create?
To create means to originate something new from intention or material.
pack-grounded (en_core_action_v1).
[PACK ] What does make mean?
To make means to bring something into existence by directed action.
pack-grounded (en_core_action_v1).
[PACK ] What is to use?
To use means to put something into service for a purpose. pack-grounded
(en_core_action_v1).
[PACK ] Define change.
To change means to make or become different in some respect. pack-grounded
(en_core_action_v1).
=== Quantitative ===
[PACK ] What does all mean?
All specifies the totality of items in a referenced set. pack-grounded
(en_core_quantitative_v1).
[PACK ] Define some.
Some specifies an indefinite nonzero subset of a referenced set.
pack-grounded (en_core_quantitative_v1).
[PACK ] What is more?
More specifies a greater quantity along a comparative scale. pack-grounded
(en_core_quantitative_v1).
[PACK ] What does enough mean?
Enough specifies a quantity sufficient to satisfy a stated requirement.
pack-grounded (en_core_quantitative_v1).
=== Causation ===
[PACK ] What is an effect?
Effect is a state of affairs produced by an antecedent cause. pack-grounded
(en_core_causation_v1).
[PACK ] Define outcome.
Outcome is the end state produced by some process. pack-grounded
(en_core_causation_v1).
[PACK ] What is a consequence?
Consequence is a state of affairs that follows necessarily from another.
pack-grounded (en_core_causation_v1).
[PACK ] What does trigger mean?
To trigger means to initiate or set in motion a subsequent event.
pack-grounded (en_core_causation_v1).
=== Polarity / frequency ===
[PACK ] What is yes?
Yes is uttered to affirm assent to a preceding question or claim.
pack-grounded (en_core_polarity_v1).
[PACK ] What does always mean?
Always indicates the frequency of occurring without exception across all
instances. pack-grounded (en_core_polarity_v1).
[PACK ] Define never.
Never indicates the frequency of occurring on zero instances. pack-grounded
(en_core_polarity_v1).
[PACK ] What is maybe?
Maybe indicates a tentative possibility without committed certainty.
pack-grounded (en_core_polarity_v1).
=== Teaching-chain (multi-clause) ===
[TEACH ] Why is truth important?
truth — teaching-grounded (cognition_chains_v1): cognition.truth; logos.core.
truth grounds knowledge (cognition.knowledge). No session evidence yet.
=== Genuinely OOV (honesty control) ===
[OOV ] What is a hypothesis?
I haven't learned 'hypothesis' yet (intent: definition). Mounted lexicon
packs: en_core_cognition_v1, en_core_meta_v1, en_core_attitude_v1,
en_core_temporal_v1, en_core_action_v1, en_core_quantitative_v1,
en_core_spatial_v1, en_core_causation_v1, en_core_polarity_v1,
en_core_relations_v1, en_core_relations_v2. Teach me via a reviewed
PackMutationProposal.
[OOV ] Define javascript.
I haven't learned 'javascript' yet (intent: definition). Mounted lexicon
packs: en_core_cognition_v1, en_core_meta_v1, en_core_attitude_v1,
en_core_temporal_v1, en_core_action_v1, en_core_quantitative_v1,
en_core_spatial_v1, en_core_causation_v1, en_core_polarity_v1,
en_core_relations_v1, en_core_relations_v2. Teach me via a reviewed
PackMutationProposal.
[OOV ] What is quasar?
I haven't learned 'quasar' yet (intent: definition). Mounted lexicon packs:
en_core_cognition_v1, en_core_meta_v1, en_core_attitude_v1,
en_core_temporal_v1, en_core_action_v1, en_core_quantitative_v1,
en_core_spatial_v1, en_core_causation_v1, en_core_polarity_v1,
en_core_relations_v1, en_core_relations_v2. Teach me via a reviewed
PackMutationProposal.
=== Cause without teaching chain (deferred SurfaceSelector target) ===
[NONE ] How does memory work?
I don't know — insufficient grounding for that yet.
[NONE ] What causes doubt?
I don't know — insufficient grounding for that yet.
======================================================================
GROUNDING DISTRIBUTION over 51 prompts:
pack 45 ( 88.2%)
oov 3 ( 5.9%)
none 2 ( 3.9%)
teaching 1 ( 2.0%)