INDEX

Explanations

addresses or address-related phrases

oai_token-act-pair · gpt-3.5-turbo

references to specific addresses

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

kus

-0.76

EY

-0.74

fps

-0.70

exper

-0.70

lime

-0.69

issance

-0.68

ISM

-0.67

uj

-0.66

 Flavoring

-0.64

fps

-0.64

POSITIVE LOGITS

 addr

0.92

 Address

0.89

 address

0.85

 spoof

0.79

 dial

0.78

 generator

0.77

address

0.77

 addresses

0.76

 verification

0.75

 Password

0.74

Activations Density 0.051%

addresses or address-related phrases

references to specific addresses

No Comments

No Known Activations