INDEX

Explanations

numbers and measurements

oai_token-act-pair · gpt-3.5-turbo

numbers and numerical references in text

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

 Vaugh

-0.77

aturdays

-0.70

 predec

-0.70

 nodd

-0.70

 traged

-0.70

 disadvant

-0.68

lambda

-0.67

 shenan

-0.67

hetti

-0.67

 neighb

-0.67

POSITIVE LOGITS

 Shin

0.66

 partName

0.63

\":

0.60

 largeDownload

0.59

 âĢº

0.57

 Spect

0.55

zbollah

0.55

Ser

0.55

 Rabb

0.53

0.52

Activations Density 0.224%

numbers and measurements

numbers and numerical references in text

No Comments

No Known Activations