INDEX

Explanations

texts related to technical specifications and manufacturing details

oai_token-act-pair · gpt-3.5-turbo

acronyms and initialisms related to organizations or political entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

selves

-0.56

 oneself

-0.49

.):

-0.49

 infringing

-0.49

 Attribution

-0.45

 anyway

-0.45

emphasis

-0.44

 anew

-0.43

 academ

-0.43

:[

-0.42

POSITIVE LOGITS

has

1.25

was

1.23

 became

1.20

 gave

1.20

 took

1.19

 went

1.16

 showed

1.15

 appeared

1.14

 seemed

1.14

 began

1.13

Activations Density 1.058%

texts related to technical specifications and manufacturing details

acronyms and initialisms related to organizations or political entities

No Comments

No Known Activations