INDEX

Explanations

phrases or words related to direct connection, interaction, or influence

oai_token-act-pair · gpt-3.5-turbo

terms related to direct actions or forms of control

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

Hide

-0.75

anners

-0.74

nesota

-0.72

tex

-0.70

 nerv

-0.68

Peb

-0.67

geist

-0.67

 Wilde

-0.66

abies

-0.66

Afee

-0.66

POSITIVE LOGITS

eur

1.01

orship

0.99

oire

0.90

 sunlight

0.89

ions

0.82

orial

0.82

orate

0.82

 descendant

0.77

ness

0.77

ivity

0.76

Activations Density 0.038%

phrases or words related to direct connection, interaction, or influence

terms related to direct actions or forms of control

No Comments

No Known Activations