INDEX

Explanations

quantitative information such as numbers, counts, and percentages

oai_token-act-pair · gpt-3.5-turbo

references to violence in media

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 manoeuv

-0.63

-)

-0.59

!'

-0.58

endif

-0.57

 instincts

-0.56

;)

-0.56

rador

-0.55

*)

-0.54

 meanwhile

-0.54

Sov

-0.54

POSITIVE LOGITS

 respective

0.63

iann

0.62

 Including

0.57

ategories

0.55

DCS

0.55

each

0.54

 respectively

0.54

excluding

0.54

 Wells

0.52

isode

0.52

Activations Density 1.605%

quantitative information such as numbers, counts, and percentages

references to violence in media

No Comments

No Known Activations