INDEX

Explanations

words related to law enforcement, specifically focusing on police-related terms

oai_token-act-pair · gpt-3.5-turbo

references to police and law enforcement issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

xual

-0.90

nces

-0.87

igible

-0.82

yip

-0.82

ceivable

-0.79

uces

-0.77

ndra

-0.74

ais

-0.73

thal

-0.71

bp

-0.71

POSITIVE LOGITS

 departments

1.16

 officers

1.09

 academy

1.06

 officer

1.02

 acad

0.99

 department

0.95

 commissioner

0.94

 Officers

0.94

 chiefs

0.94

 brutality

0.92

Activations Density 0.057%

words related to law enforcement, specifically focusing on police-related terms

references to police and law enforcement issues

No Comments

No Known Activations