INDEX

Explanations

terms related to initiating or beginning something

oai_token-act-pair · gpt-3.5-turbo

phrases related to beginnings or initiation

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

 bathrooms

-0.66

ocally

-0.64

 favors

-0.62

 magn

-0.61

 magnitude

-0.60

 impecc

-0.59

 coveted

-0.59

 attributed

-0.59

 decorated

-0.59

 loved

-0.59

POSITIVE LOGITS

start

3.83

starting

2.13

started

2.09

Start

2.09

 Start

1.93

 start

1.76

 START

1.59

starter

1.48

 Starts

1.46

begin

1.40

Activations Density 0.011%

terms related to initiating or beginning something

phrases related to beginnings or initiation

No Comments

No Known Activations