INDEX

Explanations

mentions of private entities or activities

oai_token-act-pair · gpt-3.5-turbo

terms associated with privacy

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

awa

-0.76

xual

-0.74

DAY

-0.72

 GOODMAN

-0.70

annis

-0.69

orthy

-0.67

ologies

-0.67

addons

-0.67

EFF

-0.67

amaz

-0.67

POSITIVE LOGITS

 sector

1.10

 equity

0.91

 Sector

0.82

sector

0.80

 affairs

0.78

 rented

0.78

 ownership

0.76

 jets

0.76

 property

0.75

ised

0.74

Activations Density 0.038%

mentions of private entities or activities

terms associated with privacy

No Comments

No Known Activations