INDEX

Explanations

phrases or words related to inappropriate behavior or actions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "inappropriate" and its variations in context

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

ership

-0.85

fighters

-0.78

track

-0.77

rix

-0.77

raq

-0.76

rage

-0.76

better

-0.75

glas

-0.75

ãĥĥãĥī

-0.73

reens

-0.72

POSITIVE LOGITS

 inappropriate

0.87

 behaviour

0.84

 improper

0.81

 undermin

0.80

ities

0.79

 behavior

0.79

 touching

0.79

 inappropriately

0.76

 indisc

0.74

 Parenthood

0.73

Activations Density 0.025%

phrases or words related to inappropriate behavior or actions

instances of the word "inappropriate" and its variations in context

No Comments

No Known Activations