INDEX

Explanations

references to physical locations or areas

oai_token-act-pair · gpt-3.5-turbo

references to various booths or exhibition spaces

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

hed

-0.89

itals

-0.73

 heav

-0.70

heed

-0.67

ised

-0.67

ochond

-0.65

eness

-0.65

xual

-0.65

acus

-0.64

thodox

-0.64

POSITIVE LOGITS

swer

0.85

 booth

0.77

Slot

0.74

wagon

0.69

 Tiff

0.68

Ĥª

0.67

Ħ¢

0.66

handler

0.66

ureau

0.65

igmatic

0.64

Activations Density 0.099%

references to physical locations or areas

references to various booths or exhibition spaces

No Comments

No Known Activations