INDEX

Explanations

terms related to authentic family-friendly society

oai_token-act-pair · gpt-3.5-turbo

terms related to authenticity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ware

-0.78

Ò

-0.67

 circuit

-0.67

edin

-0.66

hill

-0.62

erton

-0.62

bridge

-0.62

urden

-0.62

berus

-0.61

deck

-0.61

POSITIVE LOGITS

 different

0.81

nep

0.78

 millenn

0.78

 eleph

0.71

 volunte

0.70

hovah

0.70

 conflic

0.69

 various

0.67

ected

0.67

ogenous

0.66

Activations Density 0.000%

terms related to authentic family-friendly society

terms related to authenticity

No Comments

No Known Activations