INDEX

Explanations

phrases indicating something or someone being controlled, regulated, or influenced by external entities or rules

oai_token-act-pair · gpt-3.5-turbo

instances of the word "subject" in various contexts

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

redo

-0.52

owler

-0.52

ierra

-0.51

 Americas

-0.50

 sake

-0.50

zos

-0.49

 torches

-0.48

OUT

-0.46

irie

-0.45

oker

-0.45

POSITIVE LOGITS

to

1.23

 thereto

1.14

 unto

0.95

To

0.83

to

0.82

ivated

0.78

ivating

0.78

TO

0.75

To

0.74

ibly

0.70

Activations Density 0.107%

phrases indicating something or someone being controlled, regulated, or influenced by external entities or rules

instances of the word "subject" in various contexts

No Comments

No Known Activations