INDEX

Explanations

phrases starting with "wh" followed by a space

oai_token-act-pair · gpt-3.5-turbo

occurrences of the substring "wh"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 advance

-0.71

 Stra

-0.70

interstitial

-0.68

bed

-0.68

 Grande

-0.66

atory

-0.64

 Erdogan

-0.63

 Bach

-0.62

 Manual

-0.61

Kub

-0.61

POSITIVE LOGITS

ilst

1.22

soever

1.12

istle

1.07

ispers

1.06

olly

1.04

orf

0.97

ocom

0.95

urst

0.94

atson

0.93

irlwind

0.91

Activations Density 0.005%

phrases starting with "wh" followed by a space

occurrences of the substring "wh"

No Comments

No Known Activations