INDEX

Explanations

names related to a specific individual or company

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of a specific individual named Roh

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

 IMAGES

-0.80

iliary

-0.79

 dotted

-0.78

ciating

-0.71

 Panther

-0.65

 tongues

-0.64

 littered

-0.64

 fathers

-0.64

 Occupations

-0.63

 paycheck

-0.62

POSITIVE LOGITS

rer

0.94

Roh

0.93

mer

0.89

der

0.86

ata

0.83

rin

0.83

atche

0.83

sten

0.79

ani

0.78

ita

0.77

Activations Density 0.016%

names related to a specific individual or company

repeated mentions of a specific individual named Roh

No Comments

No Known Activations