INDEX

Explanations

documents or documentaries

oai_token-act-pair · gpt-3.5-turbo

references to documentaries

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

cffff

-0.94

hare

-0.84

hap

-0.74

Commerce

-0.72

ymes

-0.72

sembly

-0.71

ingers

-0.70

unin

-0.69

hr

-0.68

usable

-0.67

POSITIVE LOGITS

 filmmaker

1.16

 filmmaking

1.03

 documentary

1.02

 filmmakers

1.01

 film

0.95

 documentaries

0.88

arily

0.84

film

0.84

 films

0.82

arium

0.80

Activations Density 0.033%

documents or documentaries

references to documentaries

No Comments

No Known Activations