INDEX

Explanations

phrases related to "Te" in different contexts, including locations, people, and organizations

oai_token-act-pair · gpt-3.5-turbo

references to a specific entity or brand called "Te."

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

etheless

-0.72

ãģį

-0.70

ĸļ

-0.70

 Hawks

-0.69

ãĥ¼ãĥĨ

-0.68

ItemTracker

-0.68

 PRESS

-0.67

espie

-0.66

 diapers

-0.65

éĹĺ

-0.64

POSITIVE LOGITS

achable

1.31

acher

1.13

aching

1.13

levision

1.07

brate

1.05

lem

1.04

achers

1.04

lev

1.04

arers

1.02

legram

1.01

Activations Density 0.017%

phrases related to "Te" in different contexts, including locations, people, and organizations

references to a specific entity or brand called "Te."

No Comments

No Known Activations