INDEX

Explanations

occurrences of the word "Office"

oai_token-act-pair · gpt-3.5-turbo 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

oise

-0.71

udeau

-0.68

nir

-0.67

 Kaine

-0.67

ipe

-0.65

artifacts

-0.65

theless

-0.65

inging

-0.64

isers

-0.64

hered

-0.64

POSITIVE LOGITS

 Depot

0.88

0.87

gur

0.87

 Office

0.76

 Personnel

0.76

 Desk

0.73

Office

0.69

person

0.69

 Sphere

0.67

 XIII

0.66

Activations Density 0.017%

occurrences of the word "Office"

No Comments

No Known Activations