INDEX

Explanations

the word "paper."

oai_token-act-pair · gpt-3.5-turbo

references to documents or written materials, particularly academic papers

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

 Commando

-0.70

acons

-0.70

azar

-0.70

agy

-0.70

IGHTS

-0.65

 Scenes

-0.64

lement

-0.62

aton

-0.62

ooter

-0.61

eper

-0.59

POSITIVE LOGITS

 paper

3.82

paper

2.85

Paper

2.61

 Paper

2.61

 papers

2.40

papers

1.95

 parchment

1.82

 Papers

1.61

 newspaper

1.61

 manuscript

1.37

Activations Density 0.015%

the word "paper."

references to documents or written materials, particularly academic papers

No Comments

No Known Activations