INDEX

Explanations

phrases or sentences highlighting or explaining concepts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "illustrate" and its variations

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

uld

-0.76

 saline

-0.70

eworld

-0.70

ldon

-0.70

 hire

-0.68

etheless

-0.67

duct

-0.67

ught

-0.66

eat

-0.66

users

-0.65

POSITIVE LOGITS

 illustrating

1.17

 illustrates

1.06

ĸļ

0.97

 illustrate

0.94

 illustration

0.93

 Illust

0.90

 illustrated

0.87

 highlights

0.86

="#

0.86

hematic

0.84

Activations Density 0.011%

phrases or sentences highlighting or explaining concepts

instances of the word "illustrate" and its variations

No Comments

No Known Activations