INDEX

Explanations

phrases related to achievements or actions taken by individuals

oai_token-act-pair · gpt-3.5-turbo

verbs indicating actions taken or experiences had by subjects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

umbledore

-0.74

 THEN

-0.72

 Sabha

-0.71

rish

-0.67

 tomorrow

-0.64

âĢ¦]

-0.62

kHz

-0.62

OTAL

-0.62

rez

-0.62

 Malf

-0.61

POSITIVE LOGITS

 numerous

1.19

 steadily

1.16

 since

1.09

 lately

1.08

 countless

0.99

 plenty

0.95

 spor

0.95

 repeatedly

0.94

 innumerable

0.92

 consistently

0.90

Activations Density 0.411%

phrases related to achievements or actions taken by individuals

verbs indicating actions taken or experiences had by subjects

No Comments

No Known Activations