INDEX

Explanations

words related to government

oai_token-act-pair · gpt-3.5-turbo

references to government entities and actions

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

CONT

-0.78

 Mortal

-0.73

 Loop

-0.72

 Locations

-0.70

 Ancest

-0.69

Sic

-0.68

Rank

-0.67

 Duel

-0.66

 Plate

-0.65

 Warm

-0.65

POSITIVE LOGITS

 ministers

0.96

 coffers

0.94

arians

0.90

 bureaucracy

0.89

 shutdown

0.86

 agencies

0.86

 minister

0.85

 planners

0.85

asury

0.84

 ministries

0.83

Activations Density 0.053%

words related to government

references to government entities and actions

No Comments

No Known Activations