INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

references to individuals and their roles or titles

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

 needles

-0.71

 stereotypes

-0.69

Tokens

-0.65

VALUE

-0.60

 prompt

-0.59

inders

-0.58

 surrog

-0.56

Characters

-0.56

 tides

-0.56

IRO

-0.56

POSITIVE LOGITS

 retired

0.79

 retiring

0.76

KP

0.72

monds

0.71

actionDate

0.71

MD

0.68

 recovering

0.67

byn

0.66

RH

0.64

kun

0.63

Activations Density 0.511%

names of individuals

references to individuals and their roles or titles

No Comments

No Known Activations