INDEX

Explanations

references to television series

oai_token-act-pair · gpt-3.5-turbo

references to television series, particularly animated ones

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ithing

-0.75

lite

-0.74

nai

-0.74

ryn

-0.72

arching

-0.70

attled

-0.69

ovo

-0.69

imentary

-0.67

yip

-0.67

ut

-0.67

POSITIVE LOGITS

ãĥĺ

0.89

series

0.85

 finale

0.84

ĸļ

0.84

isodes

0.82

 Emin

0.81

enegger

0.80

 series

0.78

 Rollins

0.78

 è£ıè¦ļéĨĴ

0.77

Activations Density 0.022%

references to television series

references to television series, particularly animated ones

No Comments

No Known Activations