INDEX

Explanations

mentions of projects or plans

oai_token-act-pair · gpt-3.5-turbo

comma-separated lists or phrases

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

 thous

-0.65

coni

-0.63

 predec

-0.59

 Vaugh

-0.57

æ©

-0.57

 misunder

-0.56

 explan

-0.56

chuk

-0.56

 blat

-0.55

avorite

-0.54

POSITIVE LOGITS

huh

0.84

 please

0.75

 albeit

0.69

 however

0.67

eh

0.64

etc

0.64

LLC

0.63

 meanwhile

0.63

 including

0.62

 especially

0.61

Activations Density 0.261%

mentions of projects or plans

comma-separated lists or phrases

No Comments

No Known Activations