Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/elder-plinius/OBLITERATUS
/ functions
Functions
1,841 in github.com/elder-plinius/OBLITERATUS
⨍
Functions
1,841
◇
Types & classes
359
Method
test_certifier_creation
(self)
tests/test_breakthrough_modules.py:545
Method
test_certify_all_layers
(self)
tests/test_breakthrough_modules.py:620
Method
test_certify_complete_abliteration
After successful abliteration, should certify GREEN.
tests/test_breakthrough_modules.py:549
Method
test_certify_incomplete_abliteration
With clear residual refusal signal, should certify RED.
tests/test_breakthrough_modules.py:565
Method
test_ci_bounds_bracket_rate
(self)
tests/test_refusal_detection.py:119
Method
test_ci_narrower_with_more_samples
More samples should produce tighter confidence intervals.
tests/test_refusal_detection.py:138
Method
test_cka_bounded
CKA should be between 0 and 1.
tests/test_advanced_metrics.py:259
Function
test_classify_refusal_reason_tags
()
tests/test_hard_negative.py:96
Method
test_clean_elimination
After removing direction, projections should be near-zero.
tests/test_analysis.py:257
Method
test_clear_knee
Layers with a sharp dropoff should be separated by knee detection.
tests/test_abliterate.py:1674
Method
test_cluster_detection
Should detect clusters of similar directions.
tests/test_analysis.py:169
Method
test_cluster_layers_used
(self)
tests/test_informed_pipeline.py:239
Method
test_collapses_dashes
(self)
tests/test_community.py:105
Method
test_collect_activations
Test that activation collection returns correct structure.
tests/test_abliterate.py:1706
Method
test_column_count_matches_methods
(self)
tests/test_community.py:475
Method
test_combine_empty_raises
(self)
tests/test_edge_cases.py:357
Method
test_combine_single
(self)
tests/test_edge_cases.py:352
Method
test_combine_vectors
(self)
tests/test_edge_cases.py:345
Method
test_combined_mode
(self)
tests/test_advanced_metrics.py:64
Method
test_compare_base_instruct
(self)
tests/test_novel_analysis.py:287
Method
test_compare_models
(self)
tests/test_module_imports.py:58
Method
test_compare_with_alternatives
(self)
tests/test_new_analysis_modules.py:476
Method
test_compare_with_logit_lens
(self)
tests/test_new_analysis_modules.py:242
Method
test_compare_with_standard
Comparison should return valid cosine similarities.
tests/test_analysis.py:97
Method
test_component_effects_structure
(self)
tests/test_causal_and_transfer.py:120
Method
test_comprehensive_informed_report
(self)
tests/test_telemetry.py:527
Method
test_comprehensive_report
Verify that all data points are extracted from the pipeline.
tests/test_telemetry.py:427
Method
test_compute_transfer_different_models
Transfer between different models.
tests/test_breakthrough_modules.py:438
Method
test_compute_transfer_same_model
Transfer from a model to itself should have high fidelity.
tests/test_breakthrough_modules.py:419
Method
test_computes_per_expert_directions
Should compute per-expert refusal directions from routing data.
tests/test_abliterate.py:2111
Method
test_concentrated_detects_sft_or_dpo
Concentrated refusal (tail-biased) should predict SFT or DPO.
tests/test_novel_analysis.py:241
Method
test_concentration_bounded
(self)
tests/test_defense_robustness.py:82
Method
test_concept_cone_nan_direction
ConceptConeAnalyzer with NaN in activations — documenting behavior.
tests/test_edge_cases.py:76
Method
test_concept_cone_varying_prompt_counts
Concept cone with varying numbers of prompts.
tests/test_edge_cases.py:251
Method
test_concern_flagging
(self)
tests/test_refusal_detection.py:45
Method
test_cone_dimensionality_bounds
Extreme cone dimensionality values are handled.
tests/test_informed_pipeline.py:369
Method
test_config_with_zero_experts
num_local_experts=0 should not trigger MoE.
tests/test_architecture_profiles.py:568
Method
test_contribute_flag_accepted
Verify the --contribute flag parses without error.
tests/test_community.py:486
Method
test_contribute_flags_on_obliterate
Verify --contribute and --contribute-notes are accepted args.
tests/test_cli.py:122
Method
test_correlation_causal_agreement_bounded
(self)
tests/test_causal_and_transfer.py:130
Method
test_corruption_reduces_strength
(self)
tests/test_causal_and_transfer.py:103
Method
test_cosine_matrix_diagonal_ones
Diagonal of cosine matrix should be 1.0.
tests/test_analysis.py:223
Method
test_cosine_matrix_symmetry
Cosine matrix should be symmetric.
tests/test_analysis.py:214
Method
test_cosine_similarity_zero_vectors
Cosine similarity between zero vectors.
tests/test_edge_cases.py:122
Method
test_cosine_with_analytical
Learned direction should align with analytical direction.
tests/test_causal_and_transfer.py:298
Method
test_creates_output_dir
(self, tmp_path)
tests/test_community.py:201
Method
test_cross_category_similar
Similar categories should cluster together.
tests/test_causal_and_transfer.py:407
Method
test_cross_category_specificity
(self)
tests/test_causal_and_transfer.py:428
Method
test_cross_entropy_finite
(self)
tests/test_causal_and_transfer.py:363
Method
test_cross_layer
(self)
tests/test_causal_and_transfer.py:438
Method
test_cross_layer_adjacent_vs_distant
Adjacent layers typically have higher transfer than distant ones.
tests/test_causal_and_transfer.py:447
Method
test_cross_layer_empty_directions
(self)
tests/test_edge_cases.py:149
Method
test_cross_layer_single_layer
Single layer directions should still produce a result.
tests/test_edge_cases.py:505
Method
test_cross_layer_various_dims
(self, hidden_dim)
tests/test_edge_cases.py:454
Method
test_cross_model_identical
Identical directions should give perfect transfer.
tests/test_causal_and_transfer.py:375
Method
test_cross_model_random
Random directions should give low transfer.
tests/test_causal_and_transfer.py:385
Method
test_cross_model_structure
(self)
tests/test_causal_and_transfer.py:397
Method
test_cumulative_reduction_monotonic
(self)
tests/test_new_analysis_modules.py:408
Method
test_curvature_estimation
(self)
tests/test_breakthrough_modules.py:132
Method
test_custom_component_types
(self)
tests/test_causal_and_transfer.py:146
Method
test_custom_flags
(self)
tests/test_informed_pipeline.py:127
Method
test_custom_metric
(self)
tests/test_community.py:463
Method
test_custom_metric
(self)
tests/test_new_analysis_modules.py:334
Method
test_custom_prompts
(self)
tests/test_abliterate.py:165
Method
test_custom_token_texts
(self)
tests/test_novel_analysis.py:401
Method
test_dataclass_fields
(self)
tests/test_heretic_eval.py:196
Method
test_decay_rate_positive
(self)
tests/test_novel_analysis.py:382
Method
test_deepseek_v3_is_large_moe
(self)
tests/test_architecture_profiles.py:82
Method
test_default_analysis_flags
(self, pipeline)
tests/test_informed_pipeline.py:106
Method
test_default_category_map
(self)
tests/test_novel_analysis.py:196
Method
test_default_method
Default method should be advanced.
tests/test_abliterate.py:1937
Method
test_default_method_is_advanced
(self)
tests/test_abliterate.py:184
Method
test_default_new_params
(self)
tests/test_abliterate_extended.py:105
Method
test_default_prompts
(self)
tests/test_abliterate.py:160
Method
test_default_report
(self)
tests/test_informed_pipeline.py:68
Method
test_default_robustness
(self, insights)
tests/test_informed_pipeline.py:59
Function
test_default_self_improve_params_are_size_aware
()
tests/test_model_profile.py:52
Method
test_default_values
(self, insights)
tests/test_informed_pipeline.py:42
Method
test_defaults
(self)
tests/test_abliterate.py:176
Method
test_delta_calculation
(self)
tests/test_report.py:58
Method
test_dense_config
Config without MoE attributes should be dense.
tests/test_architecture_profiles.py:253
Method
test_dense_label
(self)
tests/test_architecture_profiles.py:55
Method
test_dense_reasoning_full
(self)
tests/test_architecture_profiles.py:460
Method
test_dense_recommended_method
(self)
tests/test_architecture_profiles.py:59
Method
test_dense_standard_anti_ouroboros_on
Dense Standard: Anti-Ouroboros ON for self-repair mapping.
tests/test_architecture_profiles.py:301
Method
test_dense_standard_full
(self)
tests/test_architecture_profiles.py:449
Method
test_dense_standard_no_riemannian
Dense Standard: Riemannian OFF (manifolds are flat).
tests/test_architecture_profiles.py:296
Method
test_dense_standard_spectral_cert_on
Dense Standard: Spectral cert ON for verification.
tests/test_architecture_profiles.py:306
Method
test_dense_vs_sparse_comparison
Dense projection should modify all rows; sparse should modify fewer.
tests/test_novel_analysis.py:568
Method
test_deterministic
(self)
tests/test_community.py:116
Method
test_deterministic_with_seed
(self)
tests/test_refusal_detection.py:151
Method
test_different_configs_different_hashes
(self)
tests/test_community.py:122
Method
test_different_dimensions
CKA should work with different hidden dimensions.
tests/test_advanced_metrics.py:268
Method
test_different_distributions
KL divergence of different distributions should be positive.
tests/test_advanced_metrics.py:106
Method
test_different_models_positive_kl
(self, simple_models)
tests/test_heretic_eval.py:165
Method
test_different_seeds_differ
Different seeds should produce different tensors.
tests/test_abliteration_math.py:290
Method
test_dimension_mismatch_handled
Cross-model with different hidden dims should truncate.
tests/test_causal_and_transfer.py:511
Method
test_direction_captures_signal
Wasserstein direction should have non-trivial refusal projection.
tests/test_new_analysis_modules.py:450
Method
test_direction_norms
(self)
tests/test_telemetry.py:212
Method
test_direction_persistence
Adjacent layers with similar directions should have high persistence.
tests/test_telemetry.py:224
← previous
next →
1,101–1,200 of 1,841, ranked by callers