INDEX

Explanations

It appears there is an error; neuron 4 does not show any activations above zero, suggesting it's not finding anything in the given text, or the activation data is incomplete or incorrectly displayed

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

Ago

-0.67

 Âµ

-0.66

CI

-0.64

 Bans

-0.63

 Monstrous

-0.62

ccording

-0.61

pir

-0.60

GCC

-0.60

 Ascend

-0.60

 Scala

-0.59

POSITIVE LOGITS

idon

0.86

atically

0.78

ishly

0.77

ortium

0.76

iman

0.72

ãĥĭ

0.69

idated

0.68

oney

0.67

regon

0.66

ishy

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.