INDEX

Explanations

phrases related to paper or situations where paper is involved

oai_token-act-pair · gpt-3.5-turbo

references to paper-related materials and products

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

cffffcc

-0.78

ivals

-0.75

akening

-0.72

oise

-0.71

CVE

-0.71

alez

-0.70

cius

-0.68

eston

-0.67

ostic

-0.67

aren

-0.66

POSITIVE LOGITS

clip

1.28

 towels

1.19

 towel

1.01

clips

0.98

backs

0.87

maid

0.86

Paper

0.86

pus

0.84

brush

0.82

pee

0.81

Activations Density 0.020%

phrases related to paper or situations where paper is involved

references to paper-related materials and products

No Comments

No Known Activations