INDEX

Explanations

mathematical expressions or variations in numerical values

oai_token-act-pair · gpt-3.5-turbo

statistical data represented in numerical formats

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

worm

-0.75

aline

-0.75

vertisement

-0.73

ikuman

-0.71

 stocking

-0.68

ãĥĩãĤ£

-0.68

issa

-0.68

TOM

-0.66

reads

-0.66

 Dunham

-0.65

POSITIVE LOGITS

 Sorce

0.75

#$

0.75

rawdownloadcloneembedreportprint

0.72

Ov

0.71

âĢ¢âĢ¢

0.71

 billions

0.71

 Sins

0.68

 guts

0.67

 professions

0.67

 trillions

0.64

Activations Density 0.000%

mathematical expressions or variations in numerical values

statistical data represented in numerical formats

No Comments

No Known Activations