INDEX

Explanations

names and titles of professionals

oai_token-act-pair · gpt-3.5-turbo

recurring suffix patterns in names and terms related to people

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 shorth

-0.69

 overwhelming

-0.66

Tokens

-0.63

CLS

-0.63

FIX

-0.59

 trapping

-0.57

 obsolete

-0.57

Scotland

-0.57

 apex

-0.56

 derogatory

-0.56

POSITIVE LOGITS

oglu

1.18

oulos

1.13

ewski

1.11

opoulos

1.10

icz

1.10

zyk

1.08

owski

1.04

Jr

1.00

uez

0.99

quez

0.96

Activations Density 0.273%

names and titles of professionals

recurring suffix patterns in names and terms related to people

No Comments

No Known Activations