INDEX

Explanations

words related to physical movement, specifically walking

oai_token-act-pair · gpt-3.5-turbo

references to the action of 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

ãĤ´ãĥ³

-0.80

ASE

-0.71

agate

-0.64

powder

-0.62

rim

-0.61

hap

-0.60

ymm

-0.59

ãĥł

-0.58

antic

-0.58

final

-0.57

POSITIVE LOGITS

 walks

3.75

 walk

2.01

 stroll

1.83

 walked

1.73

 walking

1.58

 Walk

1.57

 hikes

1.53

Walk

1.49

walk

1.43

 marches

1.37

Activations Density 0.011%

words related to physical movement, specifically walking

references to the action of walking

No Comments

No Known Activations