INDEX

Explanations

words related to handles or levers, particularly those associated with equipment or devices

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "lers" in various contexts

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

cum

-0.64

Cum

-0.63

 Union

-0.59

SD

-0.58

Ver

-0.57

 pulse

-0.56

UNCH

-0.55

 option

-0.55

 Dark

-0.55

 conv

-0.55

POSITIVE LOGITS

lers

4.52

ler

2.89

lings

1.89

LER

1.88

ling

1.77

lest

1.63

les

1.42

chers

1.37

tons

1.37

erers

1.36

Activations Density 0.005%

words related to handles or levers, particularly those associated with equipment or devices

repeated mentions of the word "lers" in various contexts

No Comments

No Known Activations