INDEX

Explanations

words related to starting or initiating an action

oai_token-act-pair · gpt-3.5-turbo

instances of the word "begun" or similar words indicating the start of an action or event

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 comparison

-0.61

âĺ

-0.56

oof

-0.56

lb

-0.56

RF

-0.55

 exception

-0.54

Nug

-0.54

ity

-0.54

 figure

-0.54

bies

-0.54

POSITIVE LOGITS

 begun

3.37

 commenced

1.98

 began

1.76

 started

1.74

 begin

1.65

 undertaken

1.51

 begins

1.50

 Started

1.46

started

1.42

 resumed

1.40

Activations Density 0.011%

words related to starting or initiating an action

instances of the word "begun" or similar words indicating the start of an action or event

No Comments

No Known Activations