INDEX

Explanations

references to different cultures, whether it's in the form of words related to culture or specific cultural practices

oai_token-act-pair · gpt-3.5-turbo

terms related to cults and cultures

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

deen

-0.75

 bidding

-0.70

 loophole

-0.70

 restitution

-0.69

etsk

-0.67

 loopholes

-0.65

 preferential

-0.63

 giveaways

-0.63

*/(

-0.63

 secrecy

-0.62

POSITIVE LOGITS

ivation

1.34

ivating

1.21

ivated

1.08

ures

1.06

iv

0.97

cult

0.96

urally

0.95

inct

0.94

otypes

0.88

rait

0.87

Activations Density 0.012%

references to different cultures, whether it's in the form of words related to culture or specific cultural practices

terms related to cults and cultures

No Comments

No Known Activations