INDEX

Explanations

references to government agencies or departments, particularly involving security

oai_token-act-pair · gpt-3.5-turbo

references to the "Secret Service."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 brim

-0.81

merce

-0.76

odcast

-0.66

 proportional

-0.65

tics

-0.64

 stead

-0.63

itsch

-0.62

upon

-0.62

 Springer

-0.62

 swift

-0.62

POSITIVE LOGITS

Secret

1.26

 Secret

1.12

ariat

1.10

secret

0.97

 Secrets

0.94

marine

0.93

rets

0.92

eteria

0.89

ulously

0.86

buster

0.83

Activations Density 0.007%

references to government agencies or departments, particularly involving security

references to the "Secret Service."

No Comments

No Known Activations