INDEX

Explanations

instances of the word "ear" combined with a number indicating prominence or importance

oai_token-act-pair · gpt-3.5-turbo

mentions of ears or auditory-related terms

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

ERO

-0.78

ABE

-0.67

enance

-0.67

antine

-0.66

EEK

-0.66

 Standing

-0.64

 Uniform

-0.64

CRIPTION

-0.63

 Courier

-0.63

 Vide

-0.62

POSITIVE LOGITS

ldom

1.24

 lobe

1.11

rings

1.08

marked

1.06

phones

1.06

 canal

0.98

ache

0.94

phone

0.93

ear

0.90

liest

0.89

Activations Density 0.013%

instances of the word "ear" combined with a number indicating prominence or importance

mentions of ears or auditory-related terms

No Comments

No Known Activations