INDEX

Explanations

phrases related to urging or recommending action or information

oai_token-act-pair · gpt-3.5-turbo

references to individuals or the concept of "anyone" 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 @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

adequ

-0.62

irth

-0.62

ffic

-0.61

heny

-0.60

ocamp

-0.59

ories

-0.59

bows

-0.59

pa

-0.58

itals

-0.58

Ma

-0.58

POSITIVE LOGITS

 else

1.57

THING

1.12

Else

1.12

else

1.08

 Else

1.04

soever

0.89

 doubted

0.89

body

0.89

 imaginable

0.88

0.87

Activations Density 0.019%

phrases related to urging or recommending action or information

references to individuals or the concept of "anyone" in various contexts

No Comments

No Known Activations