INDEX

Explanations

terms related to online advertisements and city names, potentially for location-based advertising

oai_token-act-pair · gpt-3.5-turbo

occurrences of advertisements and specific phrases related to media content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

cker

-0.78

hof

-0.78

holm

-0.75

agen

-0.75

ck

-0.70

gger

-0.69

mber

-0.69

ule

-0.68

 Normandy

-0.68

onomy

-0.68

POSITIVE LOGITS

 ..........

1.17

ãħĭ

1.08

ãħĭãħĭ

1.06

olulu

0.86

paio

0.86

cffffcc

0.85

taboola

0.82

aucuses

0.82

ertodd

0.81

adelphia

0.77

Activations Density 0.013%

terms related to online advertisements and city names, potentially for location-based advertising

occurrences of advertisements and specific phrases related to media content

No Comments

No Known Activations