INDEX

Explanations

references to specific television networks and shows

oai_token-act-pair · gpt-3.5-turbo

references to television networks and production companies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

htaking

-0.80

ysis

-0.71

iasis

-0.68

cipled

-0.66

 isolation

-0.65

obyl

-0.65

 Chomsky

-0.65

jri

-0.64

 Niet

-0.64

heid

-0.64

POSITIVE LOGITS

 Studios

1.08

 Films

0.93

 Presents

0.89

 Entertainment

0.86

 Corporation

0.81

 Classics

0.79

 Television

0.79

 Comics

0.79

 Records

0.79

 Publishing

0.78

Activations Density 0.092%

references to specific television networks and shows

references to television networks and production companies

No Comments

No Known Activations