INDEX

Explanations

It appears that neuron 4 does not activate for any of the provided tokens; therefore, it's not possible to determine what the neuron is looking for based on the given information

oai_token-act-pair · gpt-4-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

pei

-0.73

dev

-0.72

paralle

-0.70

PsyNetMessage

-0.70

 forever

-0.69

phys

-0.68

ãĥĺãĥ©

-0.67

DragonMagazine

-0.66

-0.65

pol

-0.64

POSITIVE LOGITS

gered

0.84

iery

0.77

backer

0.69

cing

0.65

ymes

0.64

 Coat

0.63

sts

0.62

 stocking

0.61

stood

0.61

 drainage

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.