INDEX

Explanations

words related to teasing or poking fun at someone or something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "poke" and its variations

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

ingham

-0.93

ccording

-0.71

recy

-0.69

icter

-0.68

funding

-0.64

ophers

-0.64

inance

-0.64

Leader

-0.63

effective

-0.62

Bal

-0.62

POSITIVE LOGITS

 poking

1.21

 poke

1.08

 poked

0.87

cheon

0.84

balls

0.81

ature

0.77

bowl

0.77

gery

0.76

 scratches

0.73

aly

0.71

Activations Density 0.017%

words related to teasing or poking fun at someone or something

instances of the word "poke" and its variations

No Comments

No Known Activations