INDEX

Explanations

names

oai_token-act-pair · gpt-3.5-turbo

the word "are" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ured

-0.63

isively

-0.57

ulatory

-0.56

 unsub

-0.56

obs

-0.56

shire

-0.54

ulators

-0.54

++++

-0.54

urers

-0.54

]+

-0.54

POSITIVE LOGITS

lli

1.47

tta

1.46

tto

1.40

nce

1.39

nces

1.39

ndra

1.39

nda

1.31

tsky

1.30

tti

1.27

llo

1.24

Activations Density 0.060%

names

the word "are" in various contexts

No Comments

No Known Activations