{% extends "base.html" %} {% block title_suffix %} — Sessions{% endblock %} {% block content %}

Sessions ({{ total_scored }})

{# Coverage headline (self-baseline active) #} {% if headline %}
Self-baseline active on all 5 task types. Content sessions scored vs your own lean runs: {{ headline.content_self_scored }} / {{ headline.content_sessions }} ({{ headline.content_self_pct }}%)  ·  Pre-self-baseline (B2 corpus): {{ headline.b2_content_scored }}/{{ headline.b2_content_total }} ({{ headline.b2_content_scored_pct }}%)
Empty/stub sessions (no work product): {{ headline.empty_stubs }} of {{ headline.total }} total — not scoring failures.  ·  baseline status by type →
{% endif %} {% if legacy_count %}
{{ legacy_count }} legacy session{{ "s" if legacy_count != 1 }} in this list ({{ legacy_label }}): scored before 0.15, when each response's usage was counted once per content block, so their tokens and cost are about 2x too high. They are kept and marked, and left out of baselines, tes cost, tes budget, patterns and impact. tes rescore recovers those whose transcript still exists.
{% endif %} {# Honesty element 5: relative-not-absolute framing #}
Token verdicts are relative to your own lean waste-free sessions of each type (self-baseline) where active — not an absolute efficiency standard. Baseline source shown per session. See session detail for full domain of validity.
{# Sort helpers: build href and arrow for a column header #} {%- macro sort_arrow(key) -%} {%- if sort_key == key -%} {%- if sort_dir == 'DESC' -%}↓{%- else -%}↑{%- endif -%} {%- else -%} {%- endif -%} {%- endmacro -%} {%- macro sort_href(key) -%} /?sort={{ key }}&dir={% if sort_key == key and sort_dir == 'DESC' %}asc{% else %}desc{% endif %} {%- endmacro -%} {%- macro th_sort(key, label) -%} {{ label }} {{ sort_arrow(key) }} {%- endmacro -%} {% if pairs %} {{ th_sort('date', 'Session') }} {{ th_sort('verdict', 'Token verdict') }} {{ th_sort('waste', 'Waste') }} {{ th_sort('cost', 'Cost') }} {{ th_sort('tokens', 'Tokens') }} {% for session, traj_state in pairs %} {# Honesty element 10: attribution_line as the most prominent row text #} {# Honesty elements 4+6: band badge + baseline-source badge, UNAVAILABLE neutral #} {# Trajectory — UNAVAILABLE rendered neutral (element 4) #} {% endfor %}
Trajectory
{{ session.get("attribution_line") or (session.session_id[:14] ~ "…") }} {% if session.is_legacy %} {{ legacy_label }} {% endif %} {{ session.session_id[:14] }}…  ·  {{ session.task_type }}  ·  {{ session.scored_at[:10] }} {% if session.band_verdict == "above_p75" %} above_p75 {% elif session.band_verdict == "below_p25" %} below_p25 {% elif session.band_verdict == "within_band" %} within_band {% else %} {{ session.band_verdict }} {% endif %}
{% set bs = session.get("baseline_source", "b2_corpus") %} {% if bs == "self" %} self {% elif bs == "building" %} building {% elif bs == "corpus" %} corpus {% else %} b2 {% endif %}
{% if traj_state.value == "unavailable" %} — {% elif traj_state.value == "stale" %} {{ session.judge_verdict }} (stale) {% else %} {{ session.judge_verdict }} {% endif %} {% if session.waste_event_count > 0 %} {{ session.waste_event_count }} {% else %} 0 {% endif %} {% if session.get('session_cost_usd') is not none %} {{ session.session_cost_usd | cost_display(session.unpriced_models, 3) }} {% else %} — {% endif %} {% if session.get('real_tokens') %} {{ "{:,}".format(session.real_tokens) }} {% else %} — {% endif %}
{% if price_provenance %}

{{ price_provenance }} · Cost shown in session summary — not a score.

{% endif %} {% else %}

No sessions scored yet. Run tes score <path> to score a session.

{% endif %} {% endblock %}