|
AQIT 0.1.0
|
Go to the source code of this file.
Namespaces | |
| namespace | aquin |
| namespace | aquin.compute |
| namespace | aquin.compute.find_feature |
Functions | |
| str | aquin.compute.find_feature.normalize_direction (str|None direction) |
| Path | aquin.compute.find_feature._resolve_deception_fixture () |
| str | aquin.compute.find_feature._row_text (dict[str, Any] row, *str keys) |
| tuple[list[str], list[str], dict[str, Any]] | aquin.compute.find_feature.load_deception_probes (str|Path|None path) |
| list[dict[str, Any]] | aquin.compute.find_feature._rank_features (torch.Tensor honest_acts, torch.Tensor deceptive_acts, *, int top_k, str direction="both") |
| str|None | aquin.compute.find_feature._find_feature_warning (str direction, dict[str, Any]|None chosen, *, bool persisted, str conditioning="behavior", dict[str, Any]|None behavior_meta=None) |
| str | aquin.compute.find_feature.persist_experiment_record (str model_id, str key, dict[str, Any] record, *, str|None session_id=None) |
| dict[str, Any] | aquin.compute.find_feature.run_find_feature (str model_id, *, str scorer="deception", str|Path|None prompts_path=None, int|None layer=None, str|Path|None checkpoint_path=None, int top_k=20, str direction="both", str conditioning="behavior", int max_new_tokens=64, float temperature=0.0, int benchmark_top=0, str|None persist_key=None, str|None session_id=None, Any|None openai_client=None) |
Variables | |
| aquin.compute.find_feature.HONEST_LABELS = frozenset({"honest", "truthful", "true", "control", "baseline"}) | |
| aquin.compute.find_feature.DECEPTIVE_LABELS = frozenset({"deceptive", "deception", "lie", "lying", "dishonest", "false"}) | |
| tuple | aquin.compute.find_feature.TEXT_KEYS = ("instruction", "prompt", "text", "content", "honest", "deceptive") |
| aquin.compute.find_feature.VALID_DIRECTIONS = frozenset({"both", "deceptive", "honest"}) | |