INDEX

Explanations

organizations or entities that are location-based

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a location-based affiliation or organization

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

something

-0.84

years

-0.82

current

-0.81

ushima

-0.78

Ð²

-0.78

ometers

-0.77

important

-0.77

levels

-0.77

aido

-0.77

existing

-0.76

POSITIVE LOGITS

 entrepreneur

1.09

 nonprofit

1.08

 businessman

1.04

NGO

1.02

 startup

1.00

quartered

0.96

 activist

0.96

 firm

0.96

 artist

0.95

 developer

0.93

Activations Density 0.061%

organizations or entities that are location-based

phrases that indicate a location-based affiliation or organization

No Comments

No Known Activations