INDEX

Explanations

references to flags and public displays related to social or political statements

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

ynchronous

-0.84

humans

-0.81

advant

-0.79

efficients

-0.78

 latency

-0.77

apter

-0.76

>(

-0.74

ipel

-0.74

 Centauri

-0.73

 dependency

-0.72

POSITIVE LOGITS

 erected

1.25

 vandal

1.21

 graffiti

1.20

 signage

1.16

 mural

1.11

 swast

1.09

 billboards

1.08

 billboard

1.08

 banners

1.07

 inscribed

1.07

Activations Density 0.295%

references to flags and public displays related to social or political statements

No Comments

No Known Activations