INDEX

Explanations

the word "idel" with various numerical activations, possibly indicating a specific term or entity

oai_token-act-pair · gpt-3.5-turbo

references to individuals or entities associated with a particular name

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

aternity

-0.83

alez

-0.81

soDeliveryDate

-0.79

NAT

-0.76

manuel

-0.72

umbai

-0.71

ocre

-0.68

atically

-0.67

RH

-0.66

aternal

-0.65

POSITIVE LOGITS

hammer

0.85

holder

0.84

adder

0.84

er

0.82

ham

0.80

erent

0.79

backer

0.79

erd

0.78

cloth

0.78

ularity

0.77

Activations Density 0.028%

the word "idel" with various numerical activations, possibly indicating a specific term or entity

references to individuals or entities associated with a particular name

No Comments

No Known Activations