INDEX

Explanations

the name "Shane" mentioned in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals, particularly those named Shane

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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£«

-1.11

rontal

-0.95

oral

-0.89

ental

-0.84

odore

-0.84

alg

-0.83

insula

-0.81

astics

-0.80

ymph

-0.80

anwhile

-0.80

POSITIVE LOGITS

 Shane

1.27

 Bauer

0.86

 McMahon

0.86

 Wade

0.78

 Braun

0.77

 Leone

0.76

 Raphael

0.74

 Craig

0.74

 Strauss

0.73

 Trevor

0.73

Activations Density 0.007%

the name "Shane" mentioned in various contexts

references to specific individuals, particularly those named Shane

No Comments

No Known Activations