INDEX

Explanations

phrases related to politics or government institutions

oai_token-act-pair · gpt-3.5-turbo

references to 'People' in the context of organizations, governments, or groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

RAY

-0.74

rays

-0.72

roll

-0.71

Oracle

-0.68

hematically

-0.67

ripp

-0.66

hews

-0.66

 flatt

-0.66

Rober

-0.65

tops

-0.65

POSITIVE LOGITS

 Rights

0.81

 Bureau

0.80

 Tribunal

0.79

 Matters

0.78

 Programme

0.78

 Eater

0.77

 Committee

0.77

Men

0.75

 Corps

0.75

 Responsibility

0.75

Activations Density 0.045%

phrases related to politics or government institutions

references to 'People' in the context of organizations, governments, or groups

No Comments

No Known Activations