INDEX

Explanations

legal and political terms and phrases

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

hex

-0.67

ks

-0.65

perties

-0.64

lahoma

-0.64

gy

-0.63

idae

-0.61

ge

-0.60

unia

-0.60

acity

-0.60

gew

-0.59

POSITIVE LOGITS

ING

1.75

ER

1.62

ION

1.62

LY

1.61

ERS

1.61

WOR

1.61

MENT

1.60

NESS

1.59

MAN

1.59

INGS

1.59

Activations Density 1.082%

legal and political terms and phrases

No Comments

No Known Activations