INDEX

Explanations

phrases related to organizations and official entities

oai_token-act-pair · gpt-3.5-turbo

references to the United States in various contexts

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

issance

-0.85

ezvous

-0.69

SourceFile

-0.69

asta

-0.66

istically

-0.66

ogs

-0.65

razil

-0.64

phies

-0.64

amboo

-0.63

urations

-0.63

POSITIVE LOGITS

AF

1.02

ADA

0.96

AGE

0.95

NI

0.89

SR

0.88

CG

0.87

KER

0.85

ESA

0.84

SI

0.84

SY

0.84

Activations Density 0.020%

phrases related to organizations and official entities

references to the United States in various contexts

No Comments

No Known Activations