Upload README.md with huggingface_hub

README.md

@@ -1,101 +1,58 @@
 ---
 license: mit
-task_categories:
-  - visual-question-answering
-  - image-to-text
 tags:
   - hallucination
   - vlm
   - evaluation
   - diagnostic
-  - representation-analysis
-  - subspace
-pretty_name: "Subspace Validity Suite (SVS)"
-configs:
-  - config_name: default
-    data_files: []
 viewer: false
 ---
 
-# Subspace Validity Suite (SVS) — Code & Checkpoints
+# Subspace Validity Suite (SVS)
 
-**Diagnostic toolkit and experimental checkpoints for validating claimed "visual directions" in Vision-Language Models.**
+**Diagnostic toolkit for validating "visual directions" in Vision-Language Models.**
 
-> Associated paper: "The Subspace Validity Suite: Do Visual Directions in Vision-Language Models Survive Basic Sanity Checks?" (WACV 2027, Evaluations and Datasets Track)
+Paper: "What PCA-Based Visual Directions in VLMs Actually Capture" (WACV 2027)
 
-## Key Finding
+## Installation
 
-We tested 11 methods for extracting "visual directions" from VLM hidden states across 3 architectures (LLaVA-1.5, LLaVA-Next, BLIP-2), 2 text-only backbones (Vicuna, Mistral), and 2 non-VLM transformer families (BLOOM, deepseek-coder). **Every method fails the gibberish specificity test**: random character sequences activate the claimed visual directions as strongly as visual descriptions (Gib/Vis ≈ 1.00, TOST equivalence p < 0.0001, N = 4,000).
-
-## Repository Structure
-
-### `svs/` — Toolkit Code
-- `svs_toolkit.py` — Complete SVS implementation (6 diagnostic tests)
-- `__init__.py` — Package init
-
-### `experiments/` — Experiment Scripts
-All scripts are self-contained and run on Google Colab with an L4 GPU.
+```bash
+git clone https://huggingface.co/datasets/Anonymousblind/svs-subspace-validity-suite
+cd svs-subspace-validity-suite
+pip install .
+```
 
-| Script | Experiment |
-|--------|-----------|
-| `scaled_gibberish_cpu.py` | Generate 1000 prompts per type |
-| `scaled_gibberish_gpu.py` | N=4000 gibberish test on Vicuna |
-| `scaled_gibberish_stats.py` | TOST equivalence + Cohen's d |
-| `scaled_makebreak.py` | N=500 makebreak with McNemar's test |
-| `pope_evaluation.py` | POPE evaluation (GRH validation) |
-| `vista_actual.py` | VISTA actual codebase test |
-| `pairwise_discrimination.py` | SVS Test 6: pairwise content classification |
-| `impossible_classification.py` | 36-feature classifier test |
-| `theorem_verification_v2.py` | Eigenvalue dominance verification |
-| `blip2_gibberish.py` | BLIP-2 architecture test |
-| `mistral_gibberish.py` | Mistral backbone test |
-| `generalize_test.py` | Cross-domain (BLOOM + deepseek) |
+## Quick Start
 
-### `checkpoints/` — Experimental Results
-Raw results for reviewer verification. Load with `json.load()` or `np.load()`.
+```python
+from svs import SubspaceValiditySuite
+
+svs = SubspaceValiditySuite()
+report = svs.full_report(
+    directions=your_directions,       # (k, d) numpy array
+    h_visual=visual_hidden_states,    # list of (d,) arrays
+    h_gibberish=gibberish_states,     # list of (d,) arrays
+    h_factual=factual_states,         # optional
+    h_math=math_states,               # optional
+)
+svs.print_report(report)
+```
 
-| Checkpoint | Key Result |
-|-----------|-----------|
-| `gibberish_test/gpu_checkpoint.json` | N=4000 raw alpha values (9 methods × 4 layers × 4 types) |
-| `gibberish_test/statistical_summary.json` | TOST p-values, Cohen's d, bootstrap CIs |
-| `makebreak/checkpoint.json` | N=500 per-image CHAIR results |
-| `pope/pope_checkpoint.json` | POPE accuracy (baseline vs VGCD: +0.000 difference) |
-| `vista_actual/vista_actual_checkpoint.json` | 24-config VISTA gibberish test |
-| `vista_actual/vista_actual_vectors.npz` | Extracted VISTA steering vectors |
-| `classification/pairwise_discrimination.json` | 6-pair accuracy matrix (all >95%) |
-| `theorem/theorem_results.json` | Weight matrix singular value analysis |
+## Repository Contents
 
-### `directions/` — Extracted Subspaces
-Pre-computed PCA/SVD/contrastive directions at layers [8, 16, 24, 32].
+- `svs/` — Pip-installable toolkit (6 diagnostic tests)
+- `experiments/` — All experiment scripts (Colab-ready)
+- `checkpoints/` — Raw results for reviewer verification
+- `directions/` — Extracted subspace directions
 
-| File | Contents |
-|------|----------|
-| `subspaces_9methods.npz` | 9 methods × 4 layers (PCA, image, text, contrastive, etc.) |
-| `directions_vista_nullu.npz` | VISTA + Nullu approximation directions |
-| `subspaces_contrastive.npz` | Contrastive variant directions |
+## Checkpoints
 
-## Quick Verification
+Load any checkpoint to verify paper numbers:
 
 ```python
-import json, numpy as np
-
-# Verify gibberish test: Gib/Vis ≈ 1.00 for all methods
+import json
 with open("checkpoints/gibberish_test/statistical_summary.json") as f:
     stats = json.load(f)
-for method, result in stats.items():
-    print(f"{method}: Gib/Vis={result['gv_ratio']:.2f}, "
-          f"TOST p={result['tost_p']:.4f}, d={result['cohens_d']:.3f}")
-
-# Verify POPE: VGCD has zero effect
-with open("checkpoints/pope/pope_checkpoint.json") as f:
-    pope = json.load(f)
-bl = pope.get("summary", {}).get("baseline_random", {})
-vg = pope.get("summary", {}).get("vgcd_a1.5_random", {})
-print(f"Baseline POPE: {bl.get('acc', 'N/A')}")
-print(f"VGCD POPE:     {vg.get('acc', 'N/A')}")
-print(f"Difference:    {vg.get('acc', 0) - bl.get('acc', 0):.3f}")
+for method, r in stats.items():
+    print(f"{method}: Gib/Vis={r['gv_ratio']:.2f}, d={r['cohens_d']:.3f}")
 ```
-
-## License
-
-MIT