INDEX

Explanations

terms related to prototypes, testing, and protection

oai_token-act-pair · gpt-3.5-turbo

terms related to prototypes or protective measures

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

manship

-0.84

è£ħ

-0.74

upon

-0.73

forth

-0.68

hyde

-0.66

 mund

-0.63

 dece

-0.61

fulness

-0.61

BRE

-0.61

 chilling

-0.61

POSITIVE LOGITS

ected

1.62

otyp

1.50

otypes

1.48

racted

1.44

agonist

1.43

ests

1.42

ocol

1.42

ractor

1.38

ection

1.36

ector

1.33

Activations Density 0.029%

terms related to prototypes, testing, and protection

terms related to prototypes or protective measures

No Comments

No Known Activations