INDEX

Explanations

words related to investigation or examination

oai_token-act-pair · gpt-3.5-turbo

forms of the word 'test' and related concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

senal

-0.87

£ı

-0.79

plane

-0.78

srf

-0.76

bid

-0.74

tor

-0.70

sonian

-0.70

sterdam

-0.70

ccording

-0.69

²¾

-0.69

POSITIVE LOGITS

imation

0.70

imate

0.70

ests

0.64

ãĤ¤ãĥĪ

0.64

abad

0.63

 Ebola

0.63

oys

0.62

 Grounds

0.62

esting

0.62

aceae

0.62

Activations Density 0.056%

words related to investigation or examination

forms of the word 'test' and related concepts

No Comments

No Known Activations