INDEX

Explanations

words related to literary or academic works

oai_token-act-pair · gpt-3.5-turbo

references to scholarly or influential works

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

PID

-0.80

 Bots

-0.74

 Cookie

-0.68

effect

-0.66

 Deploy

-0.66

 Predator

-0.65

 twitch

-0.64

 Generic

-0.64

 Breed

-0.63

 Mane

-0.63

POSITIVE LOGITS

ographies

1.02

ibliography

0.99

 manuscripts

0.98

uscript

0.97

 essays

0.94

ographer

0.91

published

0.91

books

0.89

ographers

0.89

archive

0.87

Activations Density 0.146%

words related to literary or academic works

references to scholarly or influential works

No Comments

No Known Activations