INDEX

Explanations

the word 'trait'

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

fman

-0.63

————————

-0.61

Pand

-0.58

ãĥķãĤ©

-0.56

NetMessage

-0.56

ppa

-0.55

meier

-0.55

ãĤ¶

-0.55

Ì¶

-0.55

obil

-0.54

POSITIVE LOGITS

 itself

0.77

 herself

0.63

 himself

0.61

 Facts

0.60

 yourself

0.56

 Himself

0.56

 Improvement

0.55

 themselves

0.55

 minus

0.54

 differs

0.53

Activations Density 1.031%

the word 'trait'

No Comments

No Known Activations