INDEX

Explanations

phrases related to media coverage and public attention

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate widespread attention or coverage

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vable

-0.78

ahime

-0.73

artifacts

-0.71

ħĭ

-0.70

Password

-0.68

 Areas

-0.68

duc

-0.67

agate

-0.67

Plan

-0.66

 Alone

-0.66

POSITIVE LOGITS

 applause

1.36

 acclaim

1.33

 criticism

1.30

 scorn

1.28

 attention

1.28

ire

1.23

 backlash

1.19

 praise

1.18

 condemnation

1.18

 ridicule

1.16

Activations Density 0.182%

phrases related to media coverage and public attention

phrases that indicate widespread attention or coverage

No Comments

No Known Activations