INDEX

Explanations

content related to media attributions and editorial labels

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

wards

-0.62

SQL

-0.58

 clitor

-0.57

 asteroids

-0.53

worms

-0.53

 Quart

-0.50

 tongues

-0.49

 Clapper

-0.48

 barg

-0.48

 Kubrick

-0.48

POSITIVE LOGITS

ontent

0.76

 IMAGES

0.75

Photo

0.75

ï

0.74

ustration

0.73

reenshot

0.70

 REUTERS

0.69

isan

0.69

ickr

0.67

CBC

0.67

Activations Density 0.190%

content related to media attributions and editorial labels

No Comments

No Known Activations