INDEX

Explanations

proper nouns or specific identifiers in a text

oai_token-act-pair · gpt-3.5-turbo

phrases related to naming or identifying people or things

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ysc

-0.73

yrinth

-0.67

 Antar

-0.66

yip

-0.66

isoft

-0.66

stead

-0.65

iasm

-0.65

cession

-0.65

awaru

-0.63

UTERS

-0.63

POSITIVE LOGITS

 names

1.09

paces

1.03

plates

1.00

names

0.85

essee

0.78

 aliases

0.78

 tags

0.77

 landmarks

0.76

gger

0.74

 personalities

0.69

Activations Density 0.056%

proper nouns or specific identifiers in a text

phrases related to naming or identifying people or things

No Comments

No Known Activations