INDEX

Explanations

words related to episodes or installments, particularly numbered ones

oai_token-act-pair · gpt-3.5-turbo

references to episodes of a series or show

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

FUL

-0.84

LESS

-0.75

Pru

-0.72

 Petty

-0.70

ãĥīãĥ©ãĤ´ãĥ³

-0.69

hips

-0.66

^^^^

-0.66

holding

-0.65

WAYS

-0.65

''''

-0.65

POSITIVE LOGITS

iphany

1.39

iscopal

1.21

hemer

1.20

isodes

1.19

istle

1.17

igen

1.03

otrop

0.99

idem

0.98

alm

0.96

iph

0.94

Activations Density 0.015%

words related to episodes or installments, particularly numbered ones

references to episodes of a series or show

No Comments

No Known Activations