INDEX

Explanations

strong verbs indicating action or movement

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

rama

-0.72

ifact

-0.62

 Reilly

-0.61

Tokens

-0.58

Ã£

-0.57

egal

-0.57

omever

-0.57

iola

-0.57

udder

-0.57

ede

-0.56

POSITIVE LOGITS

BuyableInstoreAndOnline

0.68

Joined

0.62

 humanoid

0.57

seys

0.57

AFB

0.56

 awoken

0.54

 resign

0.54

 fled

0.54

\":

0.52

 Isis

0.51

Activations Density 0.641%

strong verbs indicating action or movement

No Comments

No Known Activations