INDEX

Explanations

mentions of the word "trains"

oai_token-act-pair · gpt-3.5-turbo

mentions of trains

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

wn

-0.74

uid

-0.74

hed

-0.70

cape

-0.66

 Palm

-0.65

cus

-0.61

 wiped

-0.60

hetical

-0.60

 Herb

-0.59

kin

-0.59

POSITIVE LOGITS

 trains

3.98

 train

2.52

Train

2.10

 Train

1.98

 buses

1.96

 railways

1.94

train

1.91

 Amtrak

1.61

cars

1.55

 bikes

1.55

Activations Density 0.013%

mentions of the word "trains"

mentions of trains

No Comments

No Known Activations