INDEX

Explanations

mentions of the name "Ad" or "Adrien"

oai_token-act-pair · gpt-3.5-turbo

references to the term "Ad."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

-0.86

Ń·

-0.82

 Stronghold

-0.72

 trout

-0.69

 Ragnarok

-0.69

Jr

-0.67

 Weir

-0.66

anwhile

-0.64

 Browne

-0.64

 Kubrick

-0.64

POSITIVE LOGITS

vertis

1.25

vantage

1.24

elaide

1.24

mittedly

1.16

vertising

1.15

ventures

1.12

vance

1.12

olescent

1.10

vertisements

1.10

roit

1.08

Activations Density 0.019%

mentions of the name "Ad" or "Adrien"

references to the term "Ad."

No Comments

No Known Activations