INDEX

Explanations

phrases related to a person named "Bearden"

oai_token-act-pair · gpt-3.5-turbo

references to the term "warden" 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

psc

-0.75

awaru

-0.68

ebin

-0.66

 disparate

-0.65

ulhu

-0.65

 Tinker

-0.64

 Chimera

-0.62

orie

-0.61

EStream

-0.61

 mathemat

-0.60

POSITIVE LOGITS

arden

1.15

hill

0.87

bats

0.85

esis

0.83

alty

0.79

furt

0.77

mans

0.75

burg

0.75

velt

0.75

stein

0.74

Activations Density 0.007%

phrases related to a person named "Bearden"

references to the term "warden" and its variations

No Comments

No Known Activations