INDEX

Explanations

proper nouns or names

oai_token-act-pair · gpt-3.5-turbo

names or proper nouns typically associated with individuals

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

 Carnage

-0.72

Goo

-0.69

 Masquerade

-0.69

 Sigma

-0.68

 Somali

-0.68

 istg

-0.67

 guiActiveUnfocused

-0.67

[&

-0.67

 Miko

-0.66

 lesbians

-0.65

POSITIVE LOGITS

entin

0.80

ricks

0.77

anmar

0.77

dinand

0.76

odore

0.74

rick

0.73

rys

0.71

anda

0.71

berman

0.71

ison

0.71

Activations Density 0.290%

proper nouns or names

names or proper nouns typically associated with individuals

No Comments

No Known Activations