INDEX

Explanations

technical or theoretical concepts mentioned in a document, focusing on phrases related to theories or principles

oai_token-act-pair · gpt-3.5-turbo

concepts related to theoretical versus practical discussions

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

illin

-0.76

by

-0.66

ences

-0.63

ollar

-0.63

jab

-0.63

rients

-0.61

 ions

-0.61

resses

-0.60

orate

-0.60

in

-0.59

POSITIVE LOGITS

 guise

0.87

uthor

0.79

ividual

0.67

é¾įå

0.63

wagen

0.63

 recogn

0.62

 coer

0.57

FTWARE

0.56

 guiActiveUn

0.56

ascus

0.56

Activations Density 0.045%

technical or theoretical concepts mentioned in a document, focusing on phrases related to theories or principles

concepts related to theoretical versus practical discussions

No Comments

No Known Activations