INDEX

Explanations

mentions of things that are soft

oai_token-act-pair · gpt-3.5-turbo

terms related to softness or non-aggressive attributes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

agher

-0.76

reon

-0.74

OUGH

-0.68

 Ancients

-0.68

ulhu

-0.67

naires

-0.66

Pax

-0.64

 Emir

-0.63

 Forever

-0.62

uate

-0.62

POSITIVE LOGITS

ball

1.17

ener

1.16

ening

1.14

hearted

1.03

cover

1.02

eners

0.98

 palate

0.93

balls

0.92

ened

0.87

ens

0.86

Activations Density 0.022%

mentions of things that are soft

terms related to softness or non-aggressive attributes

No Comments

No Known Activations