INDEX

Explanations

names of specific individuals or celebrities

oai_token-act-pair · gpt-3.5-turbo

names of individuals associated with film or entertainment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

bers

-1.23

dro

-0.80

bled

-0.78

dropping

-0.78

ãĤº

-0.73

bling

-0.72

behind

-0.72

blers

-0.72

bered

-0.71

govtrack

-0.70

POSITIVE LOGITS

esson

0.89

atan

0.88

terness

0.77

 Sheen

0.77

agra

0.76

arians

0.75

stellar

0.71

alien

0.70

aples

0.69

naires

0.68

Activations Density 0.032%

names of specific individuals or celebrities

names of individuals associated with film or entertainment

No Comments

No Known Activations