INDEX

Explanations

references to well-known individuals or public figures

oai_token-act-pair · gpt-3.5-turbo

instances of the word "the" in varying capitalizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

pai

-0.81

Interstitial

-0.77

Newsletter

-0.77

hops

-0.76

merce

-0.75

omes

-0.72

Ops

-0.72

otten

-0.71

wcsstore

-0.69

ews

-0.69

POSITIVE LOGITS

 billionaire

1.22

 actor

1.20

 comedian

1.19

 outspoken

1.19

 mogul

1.16

 congressman

1.15

 businessman

1.14

 rapper

1.13

 former

1.13

 filmmaker

1.12

Activations Density 0.293%

references to well-known individuals or public figures

instances of the word "the" in varying capitalizations

No Comments

No Known Activations