INDEX

Explanations

words related to negative social behaviors, particularly harassment and violence

oai_token-act-pair · gpt-3.5-turbo

terms related to harassment and abusive behavior

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

éĹĺ

-0.97

olesterol

-0.79

ethe

-0.71

stanbul

-0.71

ernels

-0.71

 Towns

-0.70

sonian

-0.70

iets

-0.69

rient

-0.69

icts

-0.68

POSITIVE LOGITS

 accus

0.92

 harassment

0.90

 harassing

0.87

 allegations

0.86

assment

0.85

 harass

0.83

 complaints

0.82

 accusations

0.81

 tactics

0.80

 leveled

0.76

Activations Density 0.068%

words related to negative social behaviors, particularly harassment and violence

terms related to harassment and abusive behavior

No Comments

No Known Activations