INDEX

Explanations

proper nouns related to a particular individual or family name in various contexts

oai_token-act-pair · gpt-3.5-turbo

the occurrences of the name "Herman."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Lanc

-0.67

cre

-0.62

Kak

-0.62

 synthetic

-0.59

 pods

-0.59

 pseudo

-0.59

Dh

-0.59

 custody

-0.58

 display

-0.57

 isolated

-0.57

POSITIVE LOGITS

erman

4.75

erm

1.44

ert

1.35

enegger

1.28

ering

1.23

erer

1.22

erg

1.21

erb

1.16

erness

1.16

erc

1.12

Activations Density 0.009%

proper nouns related to a particular individual or family name in various contexts

the occurrences of the name "Herman."

No Comments

No Known Activations