INDEX

Explanations

references to the organization "NRA" (National Rifle Association)

oai_token-act-pair · gpt-3.5-turbo

references to the National Rifle Association (NRA)

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

atted

-0.77

hw

-0.75

enance

-0.71

esley

-0.70

dress

-0.70

kj

-0.70

hai

-0.68

 Pend

-0.67

âĵĺ

-0.65

 Flowers

-0.65

POSITIVE LOGITS

 patriarch

0.82

 firearms

0.81

NRA

0.78

istani

0.76

 mosqu

0.73

 shooters

0.73

 muzzle

0.72

 gunman

0.72

senal

0.71

 fanatic

0.70

Activations Density 0.014%

references to the organization "NRA" (National Rifle Association)

references to the National Rifle Association (NRA)

No Comments

No Known Activations