INDEX

Explanations

phrases related to actions involving transferring or moving items

oai_token-act-pair · gpt-3.5-turbo

references to objects or entities that are being acted upon or manipulated

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

 Lonely

-0.71

 Jenn

-0.67

 Polk

-0.65

 Angola

-0.63

 Roose

-0.61

 Mans

-0.60

 Herrera

-0.59

ppa

-0.59

 Kelley

-0.59

Iowa

-0.57

POSITIVE LOGITS

alian

1.22

 accordingly

0.89

atic

0.83

ividual

0.82

atically

0.74

 contained

0.71

self

0.70

alia

0.69

 numbered

0.69

 selves

0.68

Activations Density 0.260%

phrases related to actions involving transferring or moving items

references to objects or entities that are being acted upon or manipulated

No Comments

No Known Activations