INDEX

Explanations

mentions of the word "nerd" in various contexts

oai_token-act-pair · gpt-3.5-turbo

terms related to "nerd" culture and identity

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

Multiple

-0.67

achusetts

-0.67

 Cipher

-0.65

Mary

-0.63

 FISA

-0.62

 Corruption

-0.62

åĮ

-0.62

CPC

-0.61

 Traff

-0.60

-0.59

POSITIVE LOGITS

ner

0.87

pei

0.86

icer

0.86

dy

0.84

ding

0.83

fed

0.81

der

0.81

Ner

0.80

umber

0.80

linger

0.80

Activations Density 0.004%

mentions of the word "nerd" in various contexts

terms related to "nerd" culture and identity

No Comments

No Known Activations