INDEX

Explanations

names of TV networks or studios

oai_token-act-pair · gpt-3.5-turbo

references to specific television networks and shows

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

clave

-0.74

quez

-0.69

cific

-0.69

cles

-0.67

jri

-0.66

Syl

-0.66

ients

-0.65

Alz

-0.65

ggies

-0.64

 Adin

-0.64

POSITIVE LOGITS

 Films

1.40

 Television

1.28

 Studios

1.27

 Movies

1.19

Universal

1.16

 Presents

1.15

 Networks

1.14

TV

1.12

 sitcom

1.12

 Premiere

1.10

Activations Density 0.141%

names of TV networks or studios

references to specific television networks and shows

No Comments

No Known Activations