INDEX

Explanations

mentions of specific events or incidents

oai_token-act-pair · gpt-3.5-turbo

references to territorial or political occupation

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 tremend

-0.71

 Doct

-0.64

 Jess

-0.62

 additions

-0.60

 paste

-0.60

 tease

-0.59

 perspect

-0.59

 colle

-0.59

orate

-0.58

 Boss

-0.58

POSITIVE LOGITS

cknow

0.69

uble

0.68

omed

0.68

uda

0.67

vertising

0.67

ATM

0.66

AD

0.65

 CrossRef

0.65

gd

0.64

usterity

0.64

Activations Density 0.000%

mentions of specific events or incidents

references to territorial or political occupation

No Comments

No Known Activations