INDEX

Explanations

words related to witnessing, accounts, and investigations

oai_token-act-pair · gpt-3.5-turbo

terms associated with eyewitness accounts and significant witnesses in narratives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

rost

-0.70

AST

-0.68

ANK

-0.68

izoph

-0.67

udder

-0.64

 Peninsula

-0.62

WAYS

-0.62

agine

-0.62

utonium

-0.61

QUEST

-0.61

POSITIVE LOGITS

es

2.34

ively

1.33

ed

1.24

esville

1.17

esian

1.16

edIn

1.15

edly

1.11

engers

1.07

ions

1.07

esy

1.06

Activations Density 0.053%

words related to witnessing, accounts, and investigations

terms associated with eyewitness accounts and significant witnesses in narratives

No Comments

No Known Activations