INDEX

Explanations

phrases related to privilege or private matters

oai_token-act-pair · gpt-3.5-turbo

terms related to privacy and privileged information

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

ORN

-0.82

WAYS

-0.81

calling

-0.76

IELD

-0.73

 Shake

-0.72

BOX

-0.70

rers

-0.70

dry

-0.67

grass

-0.67

 Ducks

-0.67

POSITIVE LOGITS

ileged

1.44

ilege

1.36

acies

1.32

ately

1.16

urrent

1.09

atism

1.04

iless

1.04

acy

0.99

ession

0.96

aband

0.93

Activations Density 0.012%

phrases related to privilege or private matters

terms related to privacy and privileged information

No Comments

No Known Activations