INDEX

Explanations

references to specific "Hub" entities or organizational structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.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

anguage

-0.86

terday

-0.80

NetMessage

-0.76

 choir

-0.72

Downloadha

-0.72

 Âł Âł Âł Âł Âł Âł Âł Âł

-0.67

chnology

-0.66

 unprotected

-0.64

DonaldTrump

-0.64

 Samoa

-0.63

POSITIVE LOGITS

bub

1.27

bard

1.25

bies

1.09

bing

1.07

bles

1.07

lot

1.06

ris

1.05

bie

1.05

bled

1.01

lins

1.00

Activations Density 0.004%

references to specific "Hub" entities or organizational structures

No Comments

No Known Activations