INDEX

Explanations

proper nouns related to television and technology

oai_token-act-pair · gpt-3.5-turbo

references to a specific television service provider and its related content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

uyomi

-0.90

ambers

-0.87

atted

-0.82

sburgh

-0.78

atter

-0.76

isters

-0.74

arro

-0.72

opped

-0.72

hops

-0.72

ebin

-0.71

POSITIVE LOGITS

wolves

1.09

 Dire

1.02

wolf

0.97

 wolves

0.82

bone

0.77

 Wolf

0.75

bol

0.73

gence

0.73

bolt

0.73

tin

0.73

Activations Density 0.011%

proper nouns related to television and technology

references to a specific television service provider and its related content

No Comments

No Known Activations