INDEX

Explanations

titles or roles related to leadership positions

oai_token-act-pair · gpt-3.5-turbo

references to people in leadership or authoritative roles

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 tatt

-0.81

 breeze

-0.69

 depiction

-0.68

 glimps

-0.68

 hone

-0.67

 pige

-0.65

 reven

-0.65

isations

-0.64

agos

-0.64

 pear

-0.63

POSITIVE LOGITS

 Subcommittee

1.23

 committee

1.18

 subcommittee

1.11

 Appropriations

1.10

 Committee

1.08

 Legislative

1.05

 Committees

1.05

 committees

1.03

 Senate

1.02

 Judiciary

1.01

Activations Density 0.136%

titles or roles related to leadership positions

references to people in leadership or authoritative roles

No Comments

No Known Activations