INDEX

Explanations

descriptions of actions or decisions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "move" 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 @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

omial

-0.78

oola

-0.71

 sqor

-0.66

iciency

-0.65

 Barton

-0.65

Cav

-0.64

inges

-0.64

acha

-0.63

icum

-0.63

sung

-0.62

POSITIVE LOGITS

able

0.90

 toward

0.81

Motion

0.81

backs

0.80

ivism

0.78

itures

0.78

over

0.76

 towards

0.76

ments

0.75

 forward

0.75

Activations Density 0.034%

descriptions of actions or decisions

instances of the word "move" in various contexts

No Comments

No Known Activations