INDEX

Explanations

mentions of positions involving staff or leadership roles

oai_token-act-pair · gpt-3.5-turbo

references to individuals in positions of authority or government staff roles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Tokens

-0.74

same

-0.70

EEK

-0.69

ador

-0.69

Joined

-0.65

bid

-0.64

NetMessage

-0.64

aunder

-0.63

luaj

-0.63

JV

-0.63

POSITIVE LOGITS

 staff

1.38

 Staff

1.27

Staff

1.08

staff

1.08

 naval

0.78

 mission

0.77

 Mission

0.77

 operations

0.75

 Operations

0.74

operation

0.74

Activations Density 0.052%

mentions of positions involving staff or leadership roles

references to individuals in positions of authority or government staff roles

No Comments

No Known Activations