INDEX

Explanations

mention of a specific term "Sn" followed by a number

oai_token-act-pair · gpt-3.5-turbo

mentions of a specific subject or character identified by the term "Sn," likely within a narrative or thematic context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 secrecy

-0.76

xual

-0.74

 Sunni

-0.73

heid

-0.68

TAIN

-0.63

 patriarchy

-0.62

 limited

-0.61

[|

-0.61

 Sikh

-0.61

 plur

-0.60

POSITIVE LOGITS

icket

1.11

akes

1.09

agging

1.09

ook

1.07

appy

1.06

oot

1.05

igger

1.04

ugg

1.04

ipes

1.01

ipers

1.01

Activations Density 0.013%

mention of a specific term "Sn" followed by a number

mentions of a specific subject or character identified by the term "Sn," likely within a narrative or thematic context

No Comments

No Known Activations