INDEX

Explanations

connections or relationships between entities or concepts

oai_token-act-pair · gpt-3.5-turbo

phrases indicating various types of connections or relationships

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 helicop

-0.74

corn

-0.67

otos

-0.66

sburg

-0.66

usc

-0.63

disk

-0.62

stall

-0.61

VC

-0.61

DIV

-0.60

 uncomp

-0.59

POSITIVE LOGITS

 links

0.98

 connections

0.97

 ties

0.95

 connection

0.94

 link

0.86

 connecting

0.83

 resemb

0.80

links

0.78

connected

0.77

ibility

0.77

Activations Density 0.050%

connections or relationships between entities or concepts

phrases indicating various types of connections or relationships

No Comments

No Known Activations