INDEX

Explanations

mentions of the word "eros"

oai_token-act-pair · gpt-3.5-turbo

references to the concept of zero or null values

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

tub

-0.82

Desktop

-0.74

strate

-0.69

Frame

-0.68

 guiActiveUnfocused

-0.67

deck

-0.66

Oak

-0.62

ãģ®å®

-0.62

Studio

-0.61

Nice

-0.61

POSITIVE LOGITS

eros

1.26

ols

0.89

aurus

0.86

heed

0.83

idad

0.81

quin

0.80

andom

0.79

ilitary

0.79

 destro

0.78

 Aires

0.78

Activations Density 0.004%

mentions of the word "eros"

references to the concept of zero or null values

No Comments

No Known Activations