INDEX

Explanations

words related to images

oai_token-act-pair · gpt-3.5-turbo

instances of the prefix "im-" indicating negation or a reversal of meaning

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

NetMessage

-0.82

ĪĴ

-0.61

 dams

-0.60

taboola

-0.60

 brace

-0.58

 grease

-0.58

Ĭ

-0.57

ttes

-0.57

 Nicarag

-0.57

Ĥª

-0.57

POSITIVE LOGITS

enez

1.33

perial

1.31

ple

1.26

bledon

1.15

ultane

1.15

ply

1.14

inished

1.14

otor

1.13

agine

1.12

plement

1.11

Activations Density 0.040%

words related to images

instances of the prefix "im-" indicating negation or a reversal of meaning

No Comments

No Known Activations