INDEX

Explanations

names of characters in movies or TV series

oai_token-act-pair · gpt-3.5-turbo

instances of opening parentheses in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 retard

-0.87

 nutrient

-0.75

 parity

-0.73

 nuts

-0.72

 lull

-0.72

 depress

-0.71

irds

-0.70

 electr

-0.69

 hazards

-0.69

olicy

-0.68

POSITIVE LOGITS

sic

1.51

formerly

1.43

pictured

1.36

aka

1.29

?)

1.28

via

1.26

?),

1.20

who

1.19

!)

1.17

which

1.17

Activations Density 0.183%

names of characters in movies or TV series

instances of opening parentheses in text

No Comments

No Known Activations