INDEX

Explanations

references to literary works or books in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to religious texts and their titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ilitary

-0.94

asio

-0.81

orescence

-0.79

 sclerosis

-0.70

 corrosion

-0.69

ilty

-0.69

00200000

-0.64

 democracy

-0.62

adow

-0.62

undai

-0.61

POSITIVE LOGITS

stores

1.24

marks

1.18

Book

1.13

book

1.10

mark

1.09

she

1.05

store

1.02

 Book

1.02

worm

1.01

seller

0.96

Activations Density 0.019%

references to literary works or books in various contexts

references to religious texts and their titles

No Comments

No Known Activations