INDEX

Explanations

names of political figures and world events

oai_token-act-pair · gpt-3.5-turbo

names of individuals and titles related to political or public figures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Michigan

-0.70

 Chloe

-0.69

 Uncharted

-0.68

 FANTASY

-0.68

 Californ

-0.65

 Cerberus

-0.64

 Vermont

-0.64

 representation

-0.64

 California

-0.64

 Philadelphia

-0.64

POSITIVE LOGITS

azi

1.11

ulla

1.07

jri

1.07

yip

1.06

adr

1.04

ubi

1.03

llah

1.03

awi

1.01

hari

1.00

dad

0.99

Activations Density 0.241%

names of political figures and world events

names of individuals and titles related to political or public figures

No Comments

No Known Activations