INDEX

Explanations

people, positions, or entities associated with various organizations or roles

oai_token-act-pair · gpt-3.5-turbo

instances of the word "are" indicating connections or groupings among entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

cue

-0.71

oise

-0.70

vironment

-0.66

cribed

-0.66

rary

-0.65

omic

-0.64

aser

-0.64

anish

-0.63

ternity

-0.63

ocity

-0.61

POSITIVE LOGITS

 vying

1.05

 currently

0.96

 awaiting

0.94

 trademarks

0.90

 expected

0.88

 suing

0.88

 members

0.86

 fixtures

0.85

 clustered

0.85

 presently

0.85

Activations Density 0.218%

people, positions, or entities associated with various organizations or roles

instances of the word "are" indicating connections or groupings among entities

No Comments

No Known Activations