INDEX

Explanations

the word "suspect" or related terms

oai_token-act-pair · gpt-3.5-turbo

references to suspects in criminal contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

issance

-0.87

psey

-0.78

parable

-0.68

aughter

-0.67

 stead

-0.67

church

-0.67

ummer

-0.67

phies

-0.66

amping

-0.64

 Remastered

-0.64

POSITIVE LOGITS

 suspect

0.70

 suspects

0.70

 warrants

0.69

 arson

0.69

ively

0.69

FBI

0.66

 robber

0.66

 lurking

0.65

Berry

0.64

FTA

0.63

Activations Density 0.059%

the word "suspect" or related terms

references to suspects in criminal contexts

No Comments

No Known Activations