INDEX

Explanations

words starting with 'm' followed by a 9 and ending in other letters

oai_token-act-pair · gpt-3.5-turbo

terms related to measurement or mass

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

é¾įåĸļå£«

-0.76

ãĥ´ãĤ¡

-0.67

 Mechdragon

-0.67

 promot

-0.67

GOT

-0.64

 slew

-0.64

 Accountability

-0.62

 Belt

-0.62

 solicitation

-0.61

 compe

-0.61

POSITIVE LOGITS

iddling

1.39

ixtures

1.35

uddled

1.31

akings

1.27

iscal

1.26

ixtape

1.25

aturation

1.25

oust

1.21

oths

1.20

oult

1.20

Activations Density 0.024%

words starting with 'm' followed by a 9 and ending in other letters

terms related to measurement or mass

No Comments

No Known Activations