INDEX

Explanations

mention of secrets that have been discovered or are hidden

oai_token-act-pair · gpt-3.5-turbo

concepts related to secrets and discoveries

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

udeb

-0.98

ĪĴ

-0.81

erate

-0.74

phasis

-0.73

reporting

-0.70

wk

-0.68

Agg

-0.67

Zup

-0.67

bernatorial

-0.67

istrates

-0.67

POSITIVE LOGITS

 secrets

1.91

 mysteries

1.77

 hidden

1.51

 secret

1.47

 treasures

1.45

 treasure

1.33

 mysterious

1.33

 mystery

1.32

 truths

1.30

 unexpl

1.27

Activations Density 0.551%

mention of secrets that have been discovered or are hidden

concepts related to secrets and discoveries

No Comments

No Known Activations