INDEX

Explanations

phrases or names related to some form of unity or totality

oai_token-act-pair · gpt-3.5-turbo

references to inclusivity or collective identity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ĸļ

-0.72

hower

-0.72

»Ĵ

-0.69

 Citation

-0.67

EStream

-0.64

eday

-0.63

EStreamFrame

-0.60

Redditor

-0.60

Cla

-0.60

 charism

-0.59

POSITIVE LOGITS

clusive

0.89

iple

0.86

theless

0.78

ente

0.73

oots

0.72

together

0.72

ighter

0.69

ahu

0.69

foundation

0.67

 sudden

0.67

Activations Density 0.073%

phrases or names related to some form of unity or totality

references to inclusivity or collective identity

No Comments

No Known Activations