INDEX

Explanations

mentions of physical locomotion

oai_token-act-pair · gpt-3.5-turbo

references to walking

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

encies

-0.91

illet

-0.73

Products

-0.70

 Dimension

-0.70

CCC

-0.70

haps

-0.67

tains

-0.65

orate

-0.65

evin

-0.64

 quota

-0.64

POSITIVE LOGITS

 walking

3.30

walking

2.29

 Walking

2.21

 walk

2.14

 walks

1.92

 wandering

1.86

 walked

1.73

 hiking

1.70

 biking

1.65

Walk

1.63

Activations Density 0.026%

mentions of physical locomotion

references to walking

No Comments

No Known Activations