INDEX

Explanations

words related to automation or technology processes

oai_token-act-pair · gpt-3.5-turbo

references to automotive concepts or auto-related terminology

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ĸļ

-1.00

mble

-0.80

soever

-0.80

nesday

-0.79

arians

-0.78

enance

-0.77

 Flavoring

-0.71

nis

-0.71

xual

-0.70

arian

-0.70

POSITIVE LOGITS

immune

1.10

cephal

0.89

cycle

0.83

exec

0.81

jack

0.80

actively

0.77

âĹ¼

0.77

 bailout

0.74

Hot

0.73

 autos

0.73

Activations Density 0.024%

words related to automation or technology processes

references to automotive concepts or auto-related terminology

No Comments

No Known Activations