INDEX

Explanations

medical terms and related numbers

oai_token-act-pair · gpt-3.5-turbo

references to statistics, types, and aspects related to media or commercial content

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

indal

-0.75

avorite

-0.73

ãĥ¼ãĥĨ

-0.71

enegger

-0.70

oaded

-0.68

aeda

-0.68

Daddy

-0.67

abee

-0.66

iqueness

-0.65

aea

-0.64

POSITIVE LOGITS

Psy

0.62

 mosaic

0.59

 largeDownload

0.54

 translator

0.53

 interpreter

0.51

 encoding

0.51

 relay

0.51

 Mech

0.50

CM

0.50

Ph

0.50

Activations Density 0.261%

medical terms and related numbers

references to statistics, types, and aspects related to media or commercial content

No Comments

No Known Activations