INDEX

Explanations

references to train stations or train-related activities

oai_token-act-pair · gpt-3.5-turbo

mentions of trains and train-related topics

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

erenn

-0.78

FACE

-0.70

eff

-0.68

Alc

-0.65

berra

-0.64

 presumptive

-0.64

Ĥİ

-0.62

 Mortal

-0.60

null

-0.59

etheless

-0.59

POSITIVE LOGITS

Train

1.05

 train

1.03

wreck

1.01

 trains

0.97

girls

0.89

roads

0.89

tracks

0.85

 Train

0.84

ement

0.83

train

0.83

Activations Density 0.011%

references to train stations or train-related activities

mentions of trains and train-related topics

No Comments

No Known Activations