INDEX

Explanations

phrases related to military intelligence or technology

oai_token-act-pair · gpt-3.5-turbo

abbreviations, particularly those related to various metrics or identifiers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

mans

-0.77

nets

-0.76

geist

-0.75

wagen

-0.73

ding

-0.73

size

-0.72

ifiers

-0.69

deck

-0.68

rules

-0.68

space

-0.67

POSITIVE LOGITS

KE

1.26

PS

1.25

AMI

1.24

LEY

1.22

OSH

1.15

LE

1.14

BS

1.14

KER

1.10

BA

1.10

KA

1.10

Activations Density 0.045%

phrases related to military intelligence or technology

abbreviations, particularly those related to various metrics or identifiers

No Comments

No Known Activations