INDEX

Explanations

names of political figures

oai_token-act-pair · gpt-3.5-turbo

suffixes or endings of words

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

LEASE

-0.76

CLOSE

-0.76

REDACTED

-0.76

xual

-0.74

Interstitial

-0.72

Magikarp

-0.69

å¤©

-0.67

è£ħ

-0.66

eers

-0.66

 è£ı

-0.64

POSITIVE LOGITS

oslav

0.90

ado

0.86

endish

0.84

enna

0.84

bley

0.83

etooth

0.82

erson

0.82

eland

0.81

ford

0.81

sey

0.80

Activations Density 0.194%

names of political figures

suffixes or endings of words

No Comments

No Known Activations