INDEX

Explanations

phrases related to decision-making and planning

oai_token-act-pair · gpt-3.5-turbo

instances of the word "figured."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ankind

-0.73

kw

-0.67

kers

-0.66

Britain

-0.66

ewitness

-0.65

vette

-0.64

emies

-0.63

ruciating

-0.63

ifestyle

-0.63

bris

-0.62

POSITIVE LOGITS

istically

0.89

 prominently

0.85

sonian

0.80

iculty

0.76

ered

0.75

 skating

0.69

ISC

0.68

ħ

0.65

ortment

0.63

doms

0.63

Activations Density 0.009%

phrases related to decision-making and planning

instances of the word "figured."

No Comments

No Known Activations