INDEX

Explanations

instances of the word "us" being addressed in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to the collective experience or perspective of a group

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

netflix

-0.84

fect

-0.70

CPC

-0.68

JUST

-0.66

more

-0.64

ritz

-0.63

iary

-0.62

NetMessage

-0.59

served

-0.58

rogram

-0.58

POSITIVE LOGITS

selves

1.15

hers

0.93

aning

0.85

 eleph

0.85

 ourselves

0.80

alg

0.79

arily

0.77

 selves

0.75

agi

0.71

ween

0.70

Activations Density 0.052%

instances of the word "us" being addressed in various contexts

references to the collective experience or perspective of a group

No Comments

No Known Activations