INDEX

Explanations

references to the concept of "void" or emptiness in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.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

ahon

-0.68

ieri

-0.64

-0.63

ographed

-0.63

Latest

-0.62

acher

-0.59

akeru

-0.59

opter

-0.59

 Harlem

-0.59

hig

-0.58

POSITIVE LOGITS

locked

0.89

0.82

shield

0.81

lust

0.80

luster

0.79

oslav

0.76

unn

0.75

locks

0.74

fall

0.72

ing

0.72

Activations Density 0.005%

references to the concept of "void" or emptiness in various contexts

No Comments

No Known Activations