INDEX

Explanations

mentions of the name "Obama" at different strengths

oai_token-act-pair · gpt-3.5-turbo

mentions of Barack Obama

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

TOR

-0.78

auri

-0.74

ruciating

-0.71

itionally

-0.71

asus

-0.70

Cth

-0.69

asma

-0.68

âĵĺ

-0.68

ongh

-0.67

 Edinburgh

-0.67

POSITIVE LOGITS

 administration

0.94

Care

0.92

Obama

0.91

 Administration

0.87

 Obama

0.86

 presidency

0.79

 Presidential

0.79

 advisor

0.77

 appoint

0.76

omics

0.76

Activations Density 0.044%

mentions of the name "Obama" at different strengths

mentions of Barack Obama

No Comments

No Known Activations