INDEX

Explanations

names starting with 'D'

oai_token-act-pair · gpt-3.5-turbo

mentions of individuals or entities with the initial 'D'

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 friction

-0.72

ãĤ¡

-0.68

 suppl

-0.66

 disposable

-0.64

 friendly

-0.64

 compliments

-0.61

 longest

-0.60

 prett

-0.59

 diameter

-0.59

 Salvation

-0.57

POSITIVE LOGITS

yson

1.33

ixie

1.23

ickson

1.21

etermined

1.20

olph

1.20

enton

1.20

etermination

1.20

etermin

1.19

ownt

1.18

oming

1.15

Activations Density 0.041%

names starting with 'D'

mentions of individuals or entities with the initial 'D'

No Comments

No Known Activations