INDEX

Explanations

phrases related to causes or reasons for certain events or conditions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating causation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

apest

-0.67

oos

-0.65

urai

-0.64

igmat

-0.61

appro

-0.57

 Letter

-0.55

cott

-0.54

IFT

-0.53

 Malt

-0.53

talk

-0.52

POSITIVE LOGITS

by

1.39

BY

1.08

by

1.07

By

0.93

 principally

0.91

By

0.90

 chiefly

0.88

 partly

0.84

bys

0.81

 solely

0.81

Activations Density 0.146%

phrases related to causes or reasons for certain events or conditions

phrases indicating causation

No Comments

No Known Activations