INDEX

Explanations

phrases related to actions and behavior

oai_token-act-pair · gpt-3.5-turbo

references to actions and their accountability or consequences

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bid

-0.80

antha

-0.71

inately

-0.66

alg

-0.66

ļéĨĴ

-0.65

skinned

-0.65

ARC

-0.64

fortable

-0.63

ANC

-0.63

 Flavoring

-0.63

POSITIVE LOGITS

 undertaken

1.02

 actions

0.92

uate

0.90

 toward

0.90

 towards

0.89

 ACTIONS

0.85

 perpetrated

0.82

 betray

0.80

 impuls

0.79

 taken

0.78

Activations Density 0.085%

phrases related to actions and behavior

references to actions and their accountability or consequences

No Comments

No Known Activations