INDEX

Explanations

words indicating an ideal or preferred situation

oai_token-act-pair · gpt-3.5-turbo

words that express an ideal or preferred state

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

 Cosponsors

-0.79

board

-0.73

ung

-0.72

ells

-0.67

 Korea

-0.67

 Super

-0.65

punk

-0.65

bucks

-0.63

 Colleg

-0.62

dream

-0.62

POSITIVE LOGITS

 ideally

1.14

 mosqu

1.11

etheless

1.11

 suited

1.09

 compr

0.98

yip

0.98

 distingu

0.95

ĵĺ

0.90

 preferably

0.90

entimes

0.89

Activations Density 0.007%

words indicating an ideal or preferred situation

words that express an ideal or preferred state

No Comments

No Known Activations