INDEX

Explanations

references to abstaining or avoiding something

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

 Rampage

-0.67

OHN

-0.61

ģĸ

-0.61

ãĥ¼ãĥ³

-0.58

senal

-0.56

Reviewer

-0.55

 guiActiveUnfocused

-0.53

OUP

-0.53

¶ħ

-0.53

Contents

-0.52

POSITIVE LOGITS

 abst

0.78

gent

0.64

ifi

0.55

rast

0.54

ild

0.53

rom

0.52

acular

0.50

lene

0.49

sb

0.48

ril

0.48

Activations Density 0.000%

references to abstaining or avoiding something

No Comments

No Known Activations