INDEX

Explanations

words related to privacy, including different variations of the word "privacy"

oai_token-act-pair · gpt-3.5-turbo

phrases related to issues of authority and governance

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

izen

-0.73

Mars

-0.72

hak

-0.71

LIST

-0.69

JECT

-0.66

 Rove

-0.64

ractive

-0.63

sen

-0.62

STON

-0.62

Flor

-0.61

POSITIVE LOGITS

cling

0.94

acies

0.87

mson

0.85

cles

0.85

encies

0.83

ency

0.81

clinton

0.81

utics

0.79

rics

0.78

utic

0.76

Activations Density 0.036%

words related to privacy, including different variations of the word "privacy"

phrases related to issues of authority and governance

No Comments

No Known Activations