INDEX

Explanations

words related to positions or titles, especially ending with 'ons'

oai_token-act-pair · gpt-3.5-turbo

instances of a specific repetitive pattern or phrase

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

XY

-0.98

ZZ

-0.72

FORE

-0.71

Print

-0.70

TEXTURE

-0.69

lished

-0.68

RTX

-0.67

ãĤ§

-0.67

RESULTS

-0.67

âĸ¬

-0.66

POSITIVE LOGITS

ons

1.41

ensical

1.37

ervative

1.34

mith

1.27

ervatives

1.10

cale

1.09

olate

1.04

hip

1.00

ilver

1.00

terday

1.00

Activations Density 0.017%

words related to positions or titles, especially ending with 'ons'

instances of a specific repetitive pattern or phrase

No Comments

No Known Activations