INDEX

Explanations

quantitative measures related to frequency or amount

oai_token-act-pair · gpt-3.5-turbo

phrases indicating quantities or frequency related to measurements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

hess

-0.80

anamo

-0.74

stones

-0.73

arden

-0.67

ITED

-0.67

apons

-0.67

pots

-0.67

avascript

-0.66

downs

-0.64

apesh

-0.64

POSITIVE LOGITS

 capita

1.33

 month

1.26

tains

1.25

 week

1.17

 hour

1.16

 year

1.13

 person

1.13

day

1.10

malink

1.03

 kilomet

1.02

Activations Density 0.037%

quantitative measures related to frequency or amount

phrases indicating quantities or frequency related to measurements

No Comments

No Known Activations