INDEX

Explanations

written statements or press releases issued by individuals or organizations

oai_token-act-pair · gpt-3.5-turbo

mentions of formal statements or communications

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

 experimented

-0.72

apter

-0.69

Appearance

-0.69

fights

-0.69

anism

-0.68

aten

-0.68

 Asians

-0.68

antry

-0.67

lords

-0.67

igion

-0.66

POSITIVE LOGITS

 statement

1.27

 scathing

1.18

 memorandum

1.06

 reminder

1.04

 rebutt

1.03

 memo

1.02

 manifesto

1.02

 Statement

1.01

 report

1.01

 apology

0.99

Activations Density 0.154%

written statements or press releases issued by individuals or organizations

mentions of formal statements or communications

No Comments

No Known Activations