INDEX

Explanations

phrases related to locations or companies named "Clar"

oai_token-act-pair · gpt-3.5-turbo

references to the word "Clar."

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

MET

-0.83

FAULT

-0.77

ORGE

-0.76

WAYS

-0.71

Yus

-0.71

 hard

-0.70

ETF

-0.69

STEM

-0.68

HS

-0.67

ħĭ

-0.64

POSITIVE LOGITS

ifications

1.39

ifying

1.36

inet

1.32

ifier

1.22

ified

1.19

ifiers

1.19

ification

1.17

ifies

1.15

issa

1.13

ities

1.07

Activations Density 0.016%

phrases related to locations or companies named "Clar"

references to the word "Clar."

No Comments

No Known Activations