INDEX

Explanations

professionals actively engaged in some form of work or operation

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

croft

-0.76

aun

-0.71

meet

-0.70

hee

-0.69

 Flavoring

-0.69

ometric

-0.65

othy

-0.65

athy

-0.65

rection

-0.65

ochond

-0.65

POSITIVE LOGITS

 alongside

0.98

 alone

0.92

 abroad

0.89

 concurrently

0.86

 beneath

0.85

 under

0.84

 underneath

0.82

 nowhere

0.81

 illegally

0.80

 elsewhere

0.80

Activations Density 0.192%

professionals actively engaged in some form of work or operation

No Comments

No Known Activations