INDEX

Explanations

words related to receiving or subscribing to programs that offer support

oai_token-act-pair · gpt-3.5-turbo

references to support programs and subscription opportunities

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

selves

-0.58

-0.51

yards

-0.50

OVER

-0.49

 Loss

-0.49

ie

-0.48

ues

-0.47

ier

-0.47

geist

-0.47

EH

-0.45

POSITIVE LOGITS

 teasp

0.60

EStreamFrame

0.57

verbal

0.55

 incent

0.55

 glim

0.54

ãĥ¯ãĥ³

0.53

risome

0.53

 unbeliev

0.52

ĸļå£«

0.51

ãĤ¦ãĤ¹

0.51

Activations Density 0.493%

words related to receiving or subscribing to programs that offer support

references to support programs and subscription opportunities

No Comments

No Known Activations