INDEX

Explanations

sentences indicating a cause-and-effect relationship or an outcome following an action

oai_token-act-pair · gpt-3.5-turbo

references to a statement, idea, or assertion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

aths

-0.73

velop

-0.73

ickets

-0.71

adle

-0.70

amia

-0.69

tones

-0.68

utical

-0.67

ricanes

-0.66

affles

-0.66

endants

-0.66

POSITIVE LOGITS

 contrasts

0.94

 culminated

0.92

 article

0.91

 week

0.89

 discrepancy

0.89

 includes

0.88

 begs

0.87

 means

0.86

 slideshow

0.86

isn

0.85

Activations Density 0.151%

sentences indicating a cause-and-effect relationship or an outcome following an action

references to a statement, idea, or assertion

No Comments

No Known Activations