INDEX

Explanations

words related to physical barriers or obstacles, especially fences

oai_token-act-pair · gpt-3.5-turbo

references to fences

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nant

-0.72

 forth

-0.67

occ

-0.64

practice

-0.63

esta

-0.63

alg

-0.62

Nir

-0.62

ounces

-0.62

arin

-0.60

)=(

-0.60

POSITIVE LOGITS

 fence

1.50

 fences

1.24

 fencing

1.08

-+-+

0.82

 encl

0.81

 este

0.80

vine

0.78

yard

0.76

 perimeter

0.75

 gate

0.74

Activations Density 0.007%

words related to physical barriers or obstacles, especially fences

references to fences

No Comments

No Known Activations