INDEX

Explanations

instances of the word "Pad" followed by a number reflecting a level or version

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "pad" in various contexts

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

ously

-0.63

IGHTS

-0.62

ãĤ¤ãĥĪ

-0.61

 Marketable

-0.60

ews

-0.60

 shortened

-0.60

 prevailing

-0.60

DEF

-0.59

icity

-0.59

 Fundamental

-0.59

POSITIVE LOGITS

awan

1.26

het

1.05

pole

1.05

Pad

1.00

arat

0.98

asus

0.94

gery

0.92

ryn

0.88

lock

0.88

rice

0.86

Activations Density 0.014%

instances of the word "Pad" followed by a number reflecting a level or version

mentions of the word "pad" in various contexts

No Comments

No Known Activations