INDEX

Explanations

technical terms, keywords, and acronyms

oai_token-act-pair · gpt-3.5-turbo

the end-of-document token

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

emale

-0.66

accompan

-0.62

pheus

-0.61

 Appalach

-0.59

 Jagu

-0.57

 Niet

-0.56

ĳå£«

-0.56

wagen

-0.55

 equivalents

-0.54

 Highlands

-0.53

POSITIVE LOGITS

mosp

0.81

lement

0.65

 SHARES

0.58

fml

0.58

poses

0.57

imes

0.56

justice

0.56

llular

0.55

FML

0.55

hesda

0.55

Activations Density 0.129%

technical terms, keywords, and acronyms

the end-of-document token

No Comments

No Known Activations