INDEX

Explanations

political party names, such as AAP, Greens, Congress, and NDP, along with related terms

oai_token-act-pair · gpt-3.5-turbo

references to political parties and movements

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

 Yellowstone

-0.91

 Twain

-0.79

obyl

-0.78

hyde

-0.77

 Dickens

-0.71

bilt

-0.68

obar

-0.68

ilation

-0.67

 Jehovah

-0.67

 Compton

-0.66

POSITIVE LOGITS

 leader

1.22

Leader

1.16

 candidate

1.11

 stronghold

1.08

 Leader

1.05

 politician

1.01

 councillor

1.00

 caucus

0.99

MLA

0.98

 leadership

0.98

Activations Density 0.127%

political party names, such as AAP, Greens, Congress, and NDP, along with related terms

references to political parties and movements

No Comments

No Known Activations