INDEX

Explanations

proper names of individuals, organizations, and entities

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

âĵĺ

-0.79

etheless

-0.78

ctrl

-0.78

ensibly

-0.76

outer

-0.74

ometimes

-0.74

ccording

-0.72

izens

-0.72

yond

-0.71

ocally

-0.71

POSITIVE LOGITS

 Memorial

1.34

Jr

1.28

 Foundation

1.22

 Scholarship

1.13

 Institute

1.13

 Award

1.11

 Associates

1.10

 Educational

1.09

 Center

1.09

 Scholars

1.04

Activations Density 0.173%

proper names of individuals, organizations, and entities

No Comments

No Known Activations