INDEX

Explanations

words related to military operations and security

oai_token-act-pair · gpt-3.5-turbo

references to advertisements or ad-related content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ãĤ¡

-0.87

Ń·

-0.71

Ago

-0.69

angu

-0.65

iders

-0.63

ãĤ©

-0.62

ĸļ

-0.62

terday

-0.62

Pes

-0.61

 Canaver

-0.60

POSITIVE LOGITS

venture

1.05

aily

1.00

ILY

0.98

irect

0.97

ynasty

0.96

vance

0.96

DL

0.96

ventures

0.95

IER

0.95

olph

0.92

Activations Density 0.013%

words related to military operations and security

references to advertisements or ad-related content

No Comments

No Known Activations