INDEX

Explanations

mentions of criminal activities and charges

oai_token-act-pair · gpt-3.5-turbo

legal terminology related to criminal charges and offenses

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 honored

-0.86

 favors

-0.85

 labor

-0.83

 Sioux

-0.82

 Whitman

-0.82

 neighbors

-0.81

 neighbor

-0.80

aucuses

-0.77

 Neighbor

-0.75

 tumor

-0.75

POSITIVE LOGITS

 offences

1.77

 offence

1.57

 licence

1.20

 breaching

1.18

 organised

1.15

 breaches

1.13

 behaviour

1.12

 griev

1.11

 indecent

1.09

 offending

1.08

Activations Density 0.225%

mentions of criminal activities and charges

legal terminology related to criminal charges and offenses

No Comments

No Known Activations