INDEX

Explanations

phrases related to geographical locations, cities, and municipal regulations

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

etsk

-0.81

zees

-0.77

anson

-0.74

uchin

-0.73

uberty

-0.72

apego

-0.71

lyak

-0.71

igham

-0.71

etheus

-0.69

iasis

-0.69

POSITIVE LOGITS

wide

1.10

 councillor

1.06

 council

1.04

 streets

1.04

 councils

1.02

 borough

1.00

 skyline

1.00

 halls

0.98

scape

0.97

 mayors

0.97

Activations Density 3.411%

phrases related to geographical locations, cities, and municipal regulations

No Comments

No Known Activations