INDEX

Explanations

the word "universal" occurring in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to universal concepts or principles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

mosp

-0.81

--------------------------------------------------------

-0.77

HOME

-0.75

Khe

-0.74

herty

-0.73

ritz

-0.72

regon

-0.70

cano

-0.70

uden

-0.69

ħĭ

-0.68

POSITIVE LOGITS

ization

1.07

ized

1.02

ised

1.02

isation

1.01

ITY

1.00

itarian

0.99

izable

0.97

isable

0.97

ities

0.92

 acclaim

0.88

Activations Density 0.013%

the word "universal" occurring in various contexts

references to universal concepts or principles

No Comments

No Known Activations