INDEX

Explanations

numerical values or measurements

oai_token-act-pair · gpt-3.5-turbo

indicators of article or series structure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

roy

-0.71

lett

-0.63

 renew

-0.57

ä½ľ

-0.57

 expand

-0.56

 tenure

-0.55

otin

-0.55

okemon

-0.55

ode

-0.54

Ara

-0.54

POSITIVE LOGITS

zinski

0.79

aturdays

0.72

 rall

0.71

ensional

0.71

abwe

0.69

ĪĴ

0.67

urst

0.67

yrus

0.67

ertodd

0.66

emort

0.66

Activations Density 0.000%

numerical values or measurements

indicators of article or series structure

No Comments

No Known Activations