INDEX

Explanations

references to a specific individual named Scalia, likely Justice Antonin Scalia

oai_token-act-pair · gpt-3.5-turbo

references to Justice Antonin Scalia

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ku

-0.67

spr

-0.67

 ende

-0.64

 Azerb

-0.64

undai

-0.63

secut

-0.62

olesc

-0.62

retched

-0.61

 medic

-0.61

ivated

-0.61

POSITIVE LOGITS

 Scalia

1.33

 Gorsuch

1.04

 justices

1.02

 TAMADRA

0.90

ervative

0.85

 Ambro

0.82

OTUS

0.82

otle

0.81

Justice

0.81

ourke

0.76

Activations Density 0.023%

references to a specific individual named Scalia, likely Justice Antonin Scalia

references to Justice Antonin Scalia

No Comments

No Known Activations