INDEX

Explanations

proper nouns, potentially names of individuals or places

oai_token-act-pair · gpt-3.5-turbo

names of individuals and entities that appear repeatedly in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ioned

-0.82

istas

-0.77

itude

-0.75

boards

-0.72

ivity

-0.68

arella

-0.68

paw

-0.67

ingly

-0.66

patrick

-0.66

ammy

-0.65

POSITIVE LOGITS

entric

0.96

llers

0.87

Î¿

0.83

earch

0.79

Î¹

0.76

trl

0.74

 Blasio

0.73

ller

0.72

ploy

0.71

hovah

0.71

Activations Density 0.024%

proper nouns, potentially names of individuals or places

names of individuals and entities that appear repeatedly in the text

No Comments

No Known Activations