INDEX

Explanations

terms related to physical locations or settings

oai_token-act-pair · gpt-3.5-turbo

words and phrases related to weight classes and rankings, particularly in a competitive context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

etically

-0.72

es

-0.71

aries

-0.69

emi

-0.67

emic

-0.65

uers

-0.65

enne

-0.65

arial

-0.65

 veter

-0.63

uary

-0.63

POSITIVE LOGITS

bottom

1.33

berry

1.32

heads

1.31

head

1.26

weight

1.26

mere

1.26

butt

1.23

berries

1.23

puff

1.22

grass

1.21

Activations Density 0.242%

terms related to physical locations or settings

words and phrases related to weight classes and rankings, particularly in a competitive context

No Comments

No Known Activations