INDEX

Explanations

phrases related to quoting someone or attributing a statement to someone

oai_token-act-pair · gpt-3.5-turbo

phrases indicating citations or references to statements made by others

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

OND

-0.69

¬¼

-0.64

ãĥ¥

-0.64

VERTISEMENT

-0.64

earth

-0.62

Ò

-0.62

RET

-0.60

MON

-0.59

RM

-0.57

 antioxid

-0.57

POSITIVE LOGITS

Downloadha

0.74

dos

0.73

utenberg

0.68

there

0.66

we

0.65

 however

0.64

 there

0.62

 albeit

0.61

anamo

0.61

she

0.60

Activations Density 0.189%

phrases related to quoting someone or attributing a statement to someone

phrases indicating citations or references to statements made by others

No Comments

No Known Activations