INDEX

Explanations

information related to criminal activity or law enforcement

oai_token-act-pair · gpt-3.5-turbo

mentions of crime and related terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

sburg

-0.90

ding

-0.76

arians

-0.75

ounter

-0.75

 dedication

-0.74

eling

-0.71

elin

-0.71

 nont

-0.71

flu

-0.70

ithing

-0.69

POSITIVE LOGITS

 Crime

1.26

 Rating

1.07

 Crimes

0.95

 Matters

0.93

 Statistics

0.90

 Fund

0.90

 Procedure

0.89

 Disorders

0.89

 Records

0.89

 Crim

0.89

Activations Density 0.033%

information related to criminal activity or law enforcement

mentions of crime and related terminology

No Comments

No Known Activations