INDEX

Explanations

keywords related to lists and bullet points

oai_token-act-pair · gpt-3.5-turbo

the word "And" used in a variety of contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

manship

-0.82

ãģĨ

-0.74

cloth

-0.72

houses

-0.68

ses

-0.65

sense

-0.64

ABE

-0.63

BILITIES

-0.63

xes

-0.63

-0.61

POSITIVE LOGITS

romeda

1.44

rea

1.36

alus

1.08

hra

1.07

rew

0.93

ersen

0.92

onso

0.90

rogens

0.84

ean

0.83

 secondly

0.81

Activations Density 0.061%

keywords related to lists and bullet points

the word "And" used in a variety of contexts

No Comments

No Known Activations