INDEX

Explanations

words related to specific individuals, possibly names

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of individuals

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

å£«

-0.72

hands

-0.71

 Colossus

-0.71

 coaster

-0.71

é¾įå¥ĳå£«

-0.70

APS

-0.69

FACE

-0.69

raft

-0.68

wcsstore

-0.68

berries

-0.67

POSITIVE LOGITS

ovic

1.27

ovi

1.03

atory

1.02

eki

0.96

osa

0.95

uchi

0.93

anca

0.93

zai

0.92

hammad

0.90

senal

0.89

Activations Density 0.024%

words related to specific individuals, possibly names

proper nouns, particularly names of individuals

No Comments

No Known Activations