INDEX

Explanations

proper names "Shane"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Shane."

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

é¾įå¥ĳå£«

-1.15

rontal

-0.94

¥µ

-0.90

insula

-0.82

astics

-0.81

yrinth

-0.81

å¥³

-0.80

sylv

-0.79

anwhile

-0.78

awaru

-0.77

POSITIVE LOGITS

 Shane

1.03

 McMahon

0.93

 Bauer

0.85

 Greene

0.79

 Bieber

0.75

 Dixon

0.75

 Craig

0.72

 Moran

0.72

vel

0.71

 Dempsey

0.71

Activations Density 0.022%

proper names "Shane"

mentions of the name "Shane."

No Comments

No Known Activations