INDEX

Explanations

phrases related to news headlines and current events

oai_token-act-pair · gpt-3.5-turbo

acronyms and abbreviations related to criminal activities and governmental terms

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

vantage

-0.66

 derby

-0.61

Downloadha

-0.61

etsk

-0.61

onomous

-0.60

baugh

-0.60

 dstg

-0.57

tarian

-0.57

 Norn

-0.57

retty

-0.57

POSITIVE LOGITS

 EDITION

1.11

OUN

1.09

ILL

1.08

INV

1.05

IZ

1.03

ARE

1.02

IVES

1.01

ANG

1.01

OWN

1.01

LE

1.00

Activations Density 0.150%

phrases related to news headlines and current events

acronyms and abbreviations related to criminal activities and governmental terms

No Comments

No Known Activations