INDEX

Explanations

phrases related to someone has a specific action or attribute

oai_token-act-pair · gpt-3.5-turbo

the occurrence of the word "has" or its variations

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

 allowance

-0.62

 supper

-0.62

 roll

-0.60

 plug

-0.60

gr

-0.60

 ounce

-0.59

 dinner

-0.58

 foul

-0.58

 unlucky

-0.58

IELD

-0.58

POSITIVE LOGITS

has

4.12

had

2.04

Has

1.72

was

1.69

ha

1.68

have

1.48

doesn

1.41

does

1.39

is

1.35

Has

1.27

Activations Density 0.012%

phrases related to someone has a specific action or attribute

the occurrence of the word "has" or its variations

No Comments

No Known Activations