INDEX

Explanations

references to individuals, particularly names and titles related to authority figures

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

New Auto-Interp

Configuration

neuronpedia/gpt2-small__res_slefr-ajt/2-res_slefr-ajt

Prompts (Dashboard)

12,288 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

46,080

Data Type

torch.float32

Hook Point

blocks.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

apollo-research/Skylion007-openwebtext-tokenizer-gpt2

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ã±o

-0.75

 Spiel

-0.75

 Chau

-0.73

vier

-0.68

crow

-0.67

 Medic

-0.66

 Passive

-0.65

 living

-0.65

ãģį

-0.64

Fib

-0.63

POSITIVE LOGITS

selves

1.23

mite

0.87

nesium

0.86

 GOODMAN

0.85

ernels

0.84

alogue

0.82

aho

0.81

 charism

0.80

ĪĴ

0.80

ahime

0.80

Activations Density 0.046%

references to individuals, particularly names and titles related to authority figures

No Comments

No Known Activations