INDEX

Explanations

words related to legal situations and actions

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate on-screen properties or characteristics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Doodle

-0.67

--------------------------------------------------------

-0.63

ļéĨĴ

-0.63

 Archdemon

-0.60

 AUTH

-0.59

 Vaughan

-0.58

 nomine

-0.58

ulhu

-0.57

¶ħ

-0.57

 Gillespie

-0.57

POSITIVE LOGITS

demand

1.24

campus

1.21

prem

1.08

camera

1.07

screen

1.07

stage

1.06

ground

1.05

site

1.03

duty

1.03

street

1.01

Activations Density 0.026%

words related to legal situations and actions

phrases that indicate on-screen properties or characteristics

No Comments

No Known Activations