From 3c69773c326602f6392af9df0bf0f6553c350630 Mon Sep 17 00:00:00 2001 From: CaptainJack2491 Date: Fri, 8 May 2026 13:39:41 +0100 Subject: docs: viva stuff --- docs/viva/images/architecture.svg | 1 + docs/viva/images/deception_by_oversight.png | Bin 0 -> 204286 bytes docs/viva/images/evaluation_pipeline.svg | 3 +++ docs/viva/images/strategic_vs_honest.png | Bin 0 -> 210226 bytes docs/viva/images/strategic_vs_honest_study1.png | Bin 0 -> 144195 bytes docs/viva/images/study1_deception_rates.png | Bin 0 -> 167117 bytes docs/viva/images/study2_framing_bars.png | Bin 0 -> 123448 bytes docs/viva/images/v1_model_landscape.png | Bin 0 -> 210958 bytes 8 files changed, 4 insertions(+) create mode 100644 docs/viva/images/architecture.svg create mode 100644 docs/viva/images/deception_by_oversight.png create mode 100644 docs/viva/images/evaluation_pipeline.svg create mode 100644 docs/viva/images/strategic_vs_honest.png create mode 100644 docs/viva/images/strategic_vs_honest_study1.png create mode 100644 docs/viva/images/study1_deception_rates.png create mode 100644 docs/viva/images/study2_framing_bars.png create mode 100644 docs/viva/images/v1_model_landscape.png (limited to 'docs/viva/images') diff --git a/docs/viva/images/architecture.svg b/docs/viva/images/architecture.svg new file mode 100644 index 0000000..bf21244 --- /dev/null +++ b/docs/viva/images/architecture.svg @@ -0,0 +1 @@ +

Statistical & Visual Analysis

Multi-Stage Judging

Data Persistence

Experiment Orchestration

Input & Configuration

Judging Metrics

Environment Sandbox

config.yaml

ConfigLoader

Scenarios Database

Oversight Levels: low/mid/high

ExperimentRunner

Agent Logic

Virtual File System

Scenario Data Files

LLM Provider API

Interaction Logs .json

Judge Engine

Regex Checks

Blackbox LLM Judge: Output Accuracy

Glassbox LLM Judge: Reasoning Intent

Consolidated Results .csv

Statistical Analysis Engine

p-values, Odds Ratios, effect sizes

Visualization Module

Charts & Heatmaps

\ No newline at end of file diff --git a/docs/viva/images/deception_by_oversight.png b/docs/viva/images/deception_by_oversight.png new file mode 100644 index 0000000..bc6d19f Binary files /dev/null and b/docs/viva/images/deception_by_oversight.png differ diff --git a/docs/viva/images/evaluation_pipeline.svg b/docs/viva/images/evaluation_pipeline.svg new file mode 100644 index 0000000..e03bc81 --- /dev/null +++ b/docs/viva/images/evaluation_pipeline.svg @@ -0,0 +1,3 @@ + + +Blackbox JudgeDeterministic pattern matchingRegexOutput-only human-like reviewReasoning-trace-enhanced reviewBasic Deception DetectionSynergy in Evaluation PipelineStructured Deception AnalysisDeep Intent UnderstandingIdentifies omissions with reasoningUncovers subtle deception with reasoningGlassbox JudgeCatches simple omissions and framing \ No newline at end of file diff --git a/docs/viva/images/strategic_vs_honest.png b/docs/viva/images/strategic_vs_honest.png new file mode 100644 index 0000000..feda47f Binary files /dev/null and b/docs/viva/images/strategic_vs_honest.png differ diff --git a/docs/viva/images/strategic_vs_honest_study1.png b/docs/viva/images/strategic_vs_honest_study1.png new file mode 100644 index 0000000..260c7b6 Binary files /dev/null and b/docs/viva/images/strategic_vs_honest_study1.png differ diff --git a/docs/viva/images/study1_deception_rates.png b/docs/viva/images/study1_deception_rates.png new file mode 100644 index 0000000..e669089 Binary files /dev/null and b/docs/viva/images/study1_deception_rates.png differ diff --git a/docs/viva/images/study2_framing_bars.png b/docs/viva/images/study2_framing_bars.png new file mode 100644 index 0000000..5d84471 Binary files /dev/null and b/docs/viva/images/study2_framing_bars.png differ diff --git a/docs/viva/images/v1_model_landscape.png b/docs/viva/images/v1_model_landscape.png new file mode 100644 index 0000000..b4df701 Binary files /dev/null and b/docs/viva/images/v1_model_landscape.png differ -- cgit v1.2.3