INDEX

Explanations

references to collectible figures, especially those related to popular culture

oai_token-act-pair · gpt-3.5-turbo

references to various representations or characters commonly associated with certain cultural contexts

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

bey

-0.78

IFE

-0.76

 Rapids

-0.76

GBT

-0.73

rip

-0.69

ntil

-0.68

ModLoader

-0.67

roe

-0.67

artment

-0.66

vid

-0.65

POSITIVE LOGITS

 figures

1.06

 Figures

0.94

hig

0.86

figure

0.86

 figure

0.82

ets

0.80

heet

0.78

 figur

0.77

hips

0.76

 prominently

0.76

Activations Density 0.012%

references to collectible figures, especially those related to popular culture

references to various representations or characters commonly associated with certain cultural contexts

No Comments

No Known Activations