INDEX

Explanations

quantitative information such as numbers and figures

oai_token-act-pair · gpt-3.5-turbo

quantifiable measurements and statistics

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

ounded

-0.73

 unfocusedRange

-0.71

 guiActiveUn

-0.70

ouston

-0.67

assetsadobe

-0.66

erey

-0.66

skirts

-0.64

kas

-0.64

OUN

-0.63

alks

-0.63

POSITIVE LOGITS

 imaginable

1.53

 EVER

1.18

 ever

1.08

 possible

1.04

ever

0.96

 conceivable

0.96

 practicable

0.91

Ever

0.90

 shortest

0.78

 Ever

0.76

Activations Density 0.455%

quantitative information such as numbers and figures

quantifiable measurements and statistics

No Comments

No Known Activations