INDEX

Explanations

references to "The Daily Show"

oai_token-act-pair · gpt-3.5-turbo

references to "Daily" media or publications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

holes

-0.92

artisan

-0.85

bourg

-0.73

OHN

-0.73

atchewan

-0.73

Â¯

-0.72

xual

-0.71

merce

-0.71

raints

-0.70

imilar

-0.70

POSITIVE LOGITS

 Beast

0.87

 Inqu

0.84

motion

0.82

 Caller

0.81

ãĥ£

0.78

 Brief

0.75

Beast

0.74

kos

0.70

ãĥ³ãĤ¸

0.70

 Weekly

0.67

Activations Density 0.010%

references to "The Daily Show"

references to "Daily" media or publications

No Comments

No Known Activations