INDEX

Explanations

words related to inappropriate or offensive content

oai_token-act-pair · gpt-3.5-turbo

terms related to obscenity and vulgarity

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

boarding

-0.87

zig

-0.87

ctl

-0.82

starter

-0.81

pei

-0.80

iard

-0.79

TT

-0.79

backs

-0.77

onen

-0.74

woods

-0.74

POSITIVE LOGITS

 blasp

0.97

 obsc

0.96

uously

0.82

lihood

0.77

 Gutenberg

0.77

 tracts

0.75

 exhib

0.74

urized

0.73

 writ

0.72

 obscene

0.71

Activations Density 0.034%

words related to inappropriate or offensive content

terms related to obscenity and vulgarity

No Comments

No Known Activations