INDEX

Explanations

phrases mentioning or related to networks

oai_token-act-pair · gpt-3.5-turbo

instances of the word "network"

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

OPLE

-0.77

ilts

-0.74

DonaldTrump

-0.73

ugh

-0.71

 prose

-0.69

 Lama

-0.69

mond

-0.69

olkien

-0.67

orical

-0.66

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

-0.66

POSITIVE LOGITS

ed

0.97

 networks

0.91

 connectivity

0.91

eers

0.86

 neutrality

0.86

link

0.85

 operator

0.85

 Networks

0.85

 network

0.85

DAQ

0.84

Activations Density 0.046%

phrases mentioning or related to networks

instances of the word "network"

No Comments

No Known Activations