INDEX

Explanations

terms related to followers or dedicated individuals

oai_token-act-pair · gpt-3.5-turbo

references to followers of various groups or individuals

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

ces

-0.76

OUT

-0.70

 Genocide

-0.70

ced

-0.67

 circumstance

-0.67

 Somalia

-0.66

cer

-0.65

geon

-0.64

posing

-0.62

cknowled

-0.62

POSITIVE LOGITS

hip

1.31

 followers

1.17

 follower

0.95

hips

0.92

lia

0.85

 adherents

0.84

ollower

0.83

 adherent

0.81

ieve

0.78

ants

0.74

Activations Density 0.010%

terms related to followers or dedicated individuals

references to followers of various groups or individuals

No Comments

No Known Activations