INDEX

Explanations

names or terms related to organizations or people

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names and titles

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

 Benz

-0.67

ãĤ®

-0.66

 FedEx

-0.63

AAC

-0.63

 Entered

-0.62

 Adjusted

-0.60

 Baylor

-0.59

wcs

-0.59

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

-0.59

 Creed

-0.58

POSITIVE LOGITS

indal

0.75

erker

0.73

ucer

0.72

agine

0.70

ossier

0.69

ioxide

0.69

ivan

0.68

stice

0.67

enium

0.67

unia

0.66

Activations Density 0.354%

names or terms related to organizations or people

proper nouns, particularly names and titles

No Comments

No Known Activations