INDEX

Explanations

words associated with negative social judgments or actions, such as demonizing, stigmatizing, and vilifying

oai_token-act-pair · gpt-3.5-turbo

terms related to stigmatization and demonization of individuals or groups

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

chart

-0.67

acca

-0.64

negie

-0.64

UTERS

-0.64

oret

-0.61

INC

-0.59

ALT

-0.59

querque

-0.59

ource

-0.58

 explan

-0.58

POSITIVE LOGITS

imaru

1.00

enance

0.80

 slurs

0.77

 bullies

0.72

 stereotypes

0.68

 insults

0.68

 prejudice

0.67

vil

0.67

 Enemy

0.66

 hated

0.64

Activations Density 0.107%

words associated with negative social judgments or actions, such as demonizing, stigmatizing, and vilifying

terms related to stigmatization and demonization of individuals or groups

No Comments

No Known Activations