INDEX

Explanations

names or proper nouns

oai_token-act-pair · gpt-3.5-turbo

references to proper nouns, particularly names and locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

-0.77

estyles

-0.67

 Mechdragon

-0.62

ynski

-0.57

%]

-0.56

indal

-0.55

 dstg

-0.54

 blame

-0.54

NK

-0.53

KDE

-0.53

POSITIVE LOGITS

pillar

0.89

isure

0.82

heim

0.78

asia

0.76

ccording

0.76

ILA

0.75

pill

0.75

zona

0.74

hurst

0.74

wered

0.74

Activations Density 0.092%

names or proper nouns

references to proper nouns, particularly names and locations

No Comments

No Known Activations