INDEX

Explanations

locations or directions, specifically related to the north

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

 TAMADRA

-0.76

ulous

-0.73

MFT

-0.68

TED

-0.68

aminer

-0.67

 Calculator

-0.67

arov

-0.67

Filter

-0.66

Documents

-0.66

PLIC

-0.66

POSITIVE LOGITS

ward

1.37

bound

1.13

side

1.11

wards

1.06

ampton

1.05

western

1.01

west

1.00

shore

0.97

 east

0.91

 hemisphere

0.91

Activations Density 0.027%

locations or directions, specifically related to the north

No Comments

No Known Activations