INDEX

Explanations

references to space-related concepts or objects

oai_token-act-pair · gpt-3.5-turbo

references to space and the term "Lem 3."

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

hoe

-0.87

 Divide

-0.78

 Gain

-0.78

 Lauder

-0.78

 Bots

-0.74

 Beacon

-0.74

 Sutherland

-0.74

 Druid

-0.74

Bloom

-0.72

alia

-0.72

POSITIVE LOGITS

 spac

2.99

 Spac

1.82

 Parm

1.39

tim

1.36

Lem

1.16

 servic

1.15

rac

1.07

apan

1.06

aram

1.04

acet

1.04

Activations Density 0.071%

references to space-related concepts or objects

references to space and the term "Lem 3."

No Comments

No Known Activations